Thank you very much. My name is Harold Goddijn. I'm one of the founders, and I'm also the CEO. We have an exciting celebration. I think we have a great lineup of speakers. We thought it was a good idea to get together after we have sold TomTom. I think you know that we're a more focused location technology company. I think it's a good moment in time to give you an overview of what we're doing, where we are, and tell you a little bit about our aspirations and the progress we're making in executing as a family. We are a more focused location technology company with a strong balance sheet and a much simplified business model. I think TomTom is uniquely positioned to fulfill a central role in the location technology space. We are independent, we are system-agnostic. We don't compete with our customers.
We're very careful with data. We don't use data for alternative business models, but we use it only to improve our products. That's what makes us a trusted player in many segments of the location technology world. Later on, I will explain where our strengths of the business come from and how we want to exploit the big trends that are now happening in location. As a subset of location, what's happening in the automotive space, which still seems very exciting. This is Amsterdam. We are in central Amsterdam, and maps are central to everything we do. Some of our customers use those maps in a platform way and building their own applications on top of that map. We provide tools, we provide those update services to manage that data properly. The application layer is done by our customers.
An example of that are Uber, Microsoft doing steps. Apple, of course, who are using our data and our update services and tools to make maps and manage the tools themselves. We can play this role because we have invested heavily in the last five years in a very efficient map-making platform. A map-making platform is a transaction map-making platform. It's a certainly large backend, and we can continuously update that map without having to do batch processing or having to release data in file maps and so on, so forth. It's a little bit the same as computing personal software, which the industry adopted some time ago. We're doing the same for maps. Not because it's just faster and it's continuous, but it also opens up the way for automation.
That's the next phase in our map-making journey, is to automate more and more of the manual labor. If you go back 10 years ago, then you could say map-making was a very laborious exercise. It was a matter of bums on seats. More maps and better maps meant more people. There was a straight relationship between that. Increasingly, we start to dig into all sorts of data sources that are there in the public domain that we're generating ourself, and we write all sorts of software to extract maximum value of that. That allows us to get better maps, make them fresher, and reduce our cost. I think that's a very critical component. It is a way of, if you spend less money on maintaining and building the map, we can spend more on innovation and differentiating technologies and applications.
We can only do that because of our people. We understand, like all leading and successful technology companies, that you can only be as good as your people. We go to great lengths of making sure that the talent we hire, that we retain them and we train them and we advance their careers, help them to get better at what they're doing. That's not only it. We also have a organization structure that is flat, as flat as we can possibly get it. We make an effort of putting the decision-making power as low in the organization as possible. We give leaders control over the input and the output within boundaries. Of course, you need to earn that trust. You need to earn that capability.
If you do, you can go a long way within TomTom without being slowed down by red tape or all sorts of things that are non-value add. I think that's one of the reasons why people like working for us and coming to TomTom. Every year, we get about 80,000 applications from people who want to work with us. That doesn't mean it's easy to attract talent, but we are an attractive employer. I think we are enjoying very good attrition rates as well. Certainly compared to the industry averages, I think we're doing better than most. When you ask people what they like about TomTom and why they recommend other people to work for us, then the first thing that come up is culture. People like the culture.
It's an entrepreneurial, fast-moving culture where you can actually make a difference and where you can get stuff done. That is important for the type of people and the type of talent we want to attract. I think the other thing that is really important and not very well understood is that we have a strong heritage in consumer applications. We know and we've learned early on how to talk to consumers, how to develop applications that are charming, that are engaging, and that are fast, and that meet end-user requirements. We've learned that obviously in the personal navigation device market, and we are making a point to keep and hang on to that knowledge and that experience in order to test hardware, to harden hardware, but also to see what works and what doesn't work. It's a very important part of our DNA and of our heritage.
We use that knowledge and that experience to very good results in the automotive industry. That typically is a little bit behind cutting-edge technology, but are also looking for partners that understand end-user behavior and end-user products. We try to do the same thing with products that don't have an end-user interface, but where we also want to learn how our customers are interacting and using that. That is especially true for HD mapping. HD mapping is a new technology. Actually, it's a different representation of reality of the world. It's high fidelity. It's 3D. End users don't see it. It sits somewhere in the silver box in a car driving, and it's part of a larger system. In order to understand the working of that system, we've also invested in our own self-driving car, and it's parked in front of the building.
It's a fully fledged, certified level 5 self-driving car. We have built that not to build autonomous driving cars ourselves, but really get a good and deep understanding of what it takes to build those cars. What the sensors are giving you, how a map interacts with that complete system. We use this car and the test environment that comes with this also to counsel our customers in the automotive industry to tell them what works, what doesn't work, and tell them results about accuracy, reliability, and so on and so forth. Again, we're not building our own self-driving stack, but we've gone a great length of testing our own products, eating our own dog food, if you like.
Being at the front, the cutting edge of technology, has helped us in the last five years or the last 10 years, I would say, to build very strong relationships with cutting-edge companies, leading technologies, mostly on the West Coast. You see the logos of a few of them, Microsoft, Uber, Verizon, Apple, there are more. Those companies are at the front line of technological development. To be honest, they don't suffer fools easily, and it helps us to get on our toes, to work very hard to keep up the pace and learn and understand what's happening, especially on the West Coast. We can separate the hype from what's real. We can understand trends at an early stage, and have good visibility, privileged visibility on what's happening in the industry at large and in the location technology industry in particular.
It's not only that we learn from those customers, but those customers are also actively contributing to make our products better. We'll talk about that later, but from most of our customers, we get a ton of data that we can process and extract value from that data. We've also entered recently with a number of our customers, editing partnerships where they can directly edit content in the map database that will then become available to everyone, and everybody wins from that. We have a better product at lower cost. It's a very important flywheel effect that we will encourage in the years to come. There's a lot of companies who want access to location data, build location applications, and it's one of our strategic objectives to make better maps at lower cost in order to save those customers in a very efficient way.
We want to put ourselves in the midst of that evolving location ecosystem. We can do that, again, because we are non-threatening. We are independent. We don't compete with our customers. We don't use data from our customers to fuel alternative business models, and that makes us a trusted partner. As a result of that, we've also managed to expand our market share in the automotive industry, very significant over the last couple of years. We've won businesses in North America, Asia, Europe, and so on and so forth. We have a couple of very strong products, including traffic information, where in Europe we have about 80% market share and 40% market share in North America, which is growing. We've won significant deals for traffic information in North America in the last couple of years.
That is all heading in the right direction. If I look at the markets that we're operating in, it doesn't come as a surprise to you that a lot is happening now. That is exciting. It's also, I guess, of course, there's a lot of fluidity, a lot of uncertainty in those periods of transition. We feel we're very well equipped to deal with whatever that's thrown at us. That is because our technology is in very good shape, we have a strong balance sheet, and we already have leading positions in some of those markets. These are the four themes where we are playing. Obviously, in the mapping ecosystem, where we increasingly want to partner with our customers or clients, not only to generate income, but also to collect data and put ourselves in the midst of that mapping ecosystem.
I think that's a very exciting opportunity, and we're making good inroads there. Alain will tell you later how that translates in efficiency, productivity, and cost for map making. In the car is a lot happening, as you know. I don't think I tell you a secret when I say that the embedded software that is now being shipped in most cars did not quite live up to end-user expectations. The industry is understanding that, is making up for lost time, and understand that in order to keep up with the smartphone world and user expectations, things need to be different, and that means that cars will go online. That's a big trend that we're seeing, and our technology is ready to play in that trend.
What you used to see in cars was different systems and different subsystems and different buttons and different user interfaces for the radio and for the HVAC, for the anti-collision system, and for the overtaking warning and so on and so forth, is all separated and distributed through the car. That's out. What is in is a more unified, consistent, clear user interface that takes all that sensor information and all that information from the car and present that in an unambiguous UI that is easy to understand, that is not distracting, and is adding to safety. We understand that. We are working on products that we will show, and Kees van Dok, our Chief Product Officer, will show you some of the thinking that goes into these type of products and getting a unified and safe user interface in the vehicle.
The other trend that we're seeing is automated driving. Although the hype has come down a little bit and the noise has come down and expectations are now a little bit more realistic, don't make any mistake, automated driving is happening. It's here to stay. There's an enormous amount of effort being put into technology to get it to work, and we are playing our role in that system with an HD map product that is leading. We have proved that that HD map system is leading because we won two big deals from large car makers, one in Japan, one in North America. We will shortly see our first commercial product for HD mapping installed in the vehicles. Before that, there is ADAS, of course, driver assisting systems.
That train has really left the station. We already started to generate significant income from ADAS capabilities that are part of our map content. The last area where we want to play is the Maps API business. That's basically services delivered to software developers who want to location-enable their own applications. That's, for us, still a small market. We are quickly maturing our product portfolio. It's important to understand that most of the technologies we're developing now are developed in the forms of APIs. Whether they end up as an end-user-facing product in a vehicle or as a more naked API in a third-party application doesn't make that much difference. There's a big synergy between our core technology development that's taking place and being able to serve the developer world through Maps APIs. We talk about it in more detail later as well.
I'm very happy that Chris Pendleton from Microsoft is here to show you what we are doing. I want to spend a little bit of time with some proof points to show you that that strategy is really delivering business result. In our connected navigation space, electric vehicle solutions, we have announced earlier this year deals with Volkswagen, with Nissan, with FCA, and with MG, as a clear sign that there is traction in the marketplace. We are integrating our in-vehicle technology with Microsoft Connected Vehicle Platform, and in combination, we can offer the full stack of the in-car user experience and what's happening in the cloud for authentication, for data analytics and so on and so forth.
I think it's a very promising partnership that we will develop over the years to come, where we, in combination, cover the whole range of technologies to make that connected car a reality. As I said earlier, in automated driving, we won two deals, and that is, at this stage of the market development, very important to us. For the first time, we will get real user feedback from commercially available products. That helps us to understand how our products are used and harden the technology and get better based on real user feedback in the marketplace. It's also a very important proof point to other car makers that need to make a decision about the deployment of HD maps in their vehicles. There is a tendency, of course, to go with what's proven in the marketplace.
The third reason why this is really important is that it will give us, for the first time, sensor data that will come back into our system at scale. That will help us to keep those HD maps fresh, of a high quality, and eventually, it will lead to self-healing HD maps, where the whole process of change detection and fixing maps can be automated to a large degree. Maps APIs, I spoke to that as well. There's also proof that that's starting to work with our partnership with Microsoft, but also through our own API store, where we are serving software developers who want to location-enable their own application. I'm very proud of what we have achieved as a team and as a business. We've gone through multiple changes. You've seen that. I feel that we are in a very good position.
We have a strong product roadmap. We have a fantastic team and a strong balance sheet. We're generating cash. I think we are at an excellent position to capture the opportunities that will open up over the next years, both midterm, but certainly in the longer term. The independent nature of our company allows us to put our customers and end users first. Again, the data we are collecting is solely used to improve our own products and not to feed alternative business models. That is why we are a trusted partner with so many triple-A companies in the world. Throughout the rest of the day, we will elaborate on this topic. We'll go deeper on every vertical so you get a better idea of what's going on, and you can visualize the progress that we're making.
I'm very happy that some of our customers will give you an outside perspective of what we're doing and what it is, what it means to work with TomTom as a partner.
My name is Taco Titulaer. I've been with the company since the IPO in 2005. I want to take you through the financial model. We'll start with revenue, then touch on the gross margin. The next step is to look at the balance sheet, also following up on what Harold is talking about the freshness of content and continuously releasable software, and the implications that that has on our capitalization amortization practices. Look at our spend. How do we spend our money? What are the trends? Where we will spend more and where we will spend less. The next one is to look at automotive backlog. It's a new KPI that we have introduced today. Also touch on enterprise, and then conclude with the outlook. TomTom acquired Tele Atlas in 2008. We announced it in 2007.
We acquired in 2008. If you look at where we were in the first year, where we had fully integrated Tele Atlas, then a lot has changed. If you look at the revenue perspective, 10 years ago, almost 90% of our revenue was coming from hardware, and today it has flipped and more than 70% or two-third of our revenue is coming from software. Why is that important? Because software tends to be more sticky. Where we sell software, we do that with long-term contract, and we have deep integration into our customers' products. The other thing that comes with software is that it has higher gross margins, mostly, and that is also the case with TomTom.
10 years ago, our gross margin was south of 50%. Today we expect our gross margin to reach north of 70% for 2019. We expect this trend to continue. The technology that we are developing and also the product that we're selling is changing. The freshness of content has materially changed. Not only because we are able to do that, but also the requirement of our customers is changing. For the pure driver navigation, the freshness is different than where we look into the future, where you have autonomous driving, where the freshness and the accuracy of the content needs to go up significantly. It's not only the content, but also on the application layer, we are making changes. We are moving from embedded software to continuously releasable software.
That has had its effect on our capitalization practices, which we announced during our Q2 results. So we will capitalize a lot less. What we capitalize, we put shorter on our balance sheet. The longest is now six years, where it used to be beyond 10 years. We will capitalize less, and we will amortize it also quicker. There's a bit of a catch-up to be done, where that's related to the Tele Atlas acquisition, where we were still planning to amortize for another 10 years. We'll do that in an accelerated manner in the next two years. You can see that in this slide. Our intangible asset position was just shy of EUR 700 million. Through the fast amortization of close to EUR 300 million, we think that we can end up just under EUR 500 million by the end of the year.
Next year, we'll do another acceleration of amortization, so will bring us just above EUR 200. After that year, you will see a new normal of amortization and capitalization. The D&A will obviously continue to decline, and will at a certain point, find the same value as CapEx. How do we spend it? If you look at the slide on the left-hand side, you see what we are expected to spend from a cash point of view. This year, that's an increase of 17%. There's a disclaimer on this slide that we don't think this trend will continue, so don't worry. We will not increase with 17% every year. We've seen this increase. We have made the necessary steps that are required in our application layer and also in engineering.
If you zoom in more in this cash spend, this EUR 500 million plus, then roughly two-third of that is coming from R&D. The remainder, 30%, is SG&A, and then 5% for marketing. Zooming in to that R&D spend, you could divide that in the application layer that is 30% and 70% coming from content. Within content, and that is what Harold already told us this morning, there's also a trend towards less spend on the sourcing to getting the sources, processing the sources, less human touch, more automation for faster cycle times, lower latencies, and more spending on the engineering to make all that happen. Machine learning from artificial intelligence, and the possibilities that that can bring us. On the right-hand side at the bottom, again, cash spend is expected to go up in the coming years, but at a much slower pace than we have seen.
Automotive backlog. When talking to investors and our analysts, we've learned that just providing the one-year guidance and also the order intake makes it difficult to predict the future. Today, we have introduced the automotive backlog, and this will replace the order intake going forward. Today, we will announce the automotive backlog, and in February, we will give an update, and we'll do that every February when we publish our full-year results. What you can see here is the cumulative expected IFRS revenue that we will foresee in the next decade plus, and a number of caveats. When we have an award, if we have signed a contract, it almost never happens that we have committed purchasing in there. Also not volumes. This is an estimate of the automotive customers of the car sales, coupled with the take rates and the agreed pricing.
The actual results that we will present, for example, in 2021 will differ. Why will it differ? On the one hand, we will reassess the future regularly, at least once a quarter. What we see in our backlog can change. The other thing, and that is positive, obviously, is that we will add deals to the 2021 picture. That's less the case, obviously, for the H2019, because that we're almost done for the year. This is a fair estimate of where IFRS revenue will end up in 2019. If we then publish the update in February, how to interpret the update is that if you compare it with what we said today and you look at the delta, there are three trends. One is the reported revenue in between, so in H2.
The order intake, number 3 is the contract reassessment per contract related to IFRS 15. Enterprise revenue, we don't have a backlog. As already explained by Harold, this is very sticky business, long tenor. Tends to be all you can eat fixed contract based on the size and based on the value that we bring to our customers. You will see shifts. On the one hand, because lots of these customers are in the U.S. Their contracts tend to be dollar-denominated. There's a Forex angle. The other difference can be if we lose a client or if we expand or introduce new clients, what we have seen this year with Microsoft. That's also the reason why the new normal of revenue is trending towards EUR 160 for this year.
Today's revenue is the best indicator for future revenue due to the nature of the contracts. To conclude, we're looking at the outlook. To start with the revenue picture. For 2018 until 2021, we expect a CAGR of 10%, bringing us to half a billion EUR. This is location technology only, so it's excluding the revenue that will come in from consumer. Beyond that, and that will be underpinned by the presentations that you will hear today as well, we see an acceleration of growth towards 15%. 15% is the midpoint of the EUR 1.5 billion-EUR 2 billion that's displayed on this slide. On cash flow, what's important for our business, also important for our customers, is that we remain profitable. Due to all the accountings going through the P&L, the accelerated amortization, deferring revenue, and what have you, we focus on free cash flow.
Our midterm aim is to bring that to a double-digit number. At the balance sheet. Our balance sheet is very healthy. Combined with the addition of the free cash flow, we are constantly reassessing that position. Four months ago, we gave back EUR 750 million to our shareholders. We will reassess that again next year, and that can lead, amongst other things, to a share buyback.
Hi, good morning. I'm Alain De Taeye. I'm a Member of the Management Board, more importantly, I make maps. I have been making maps for over 35 years. I'm still very passionate about it, especially because the evolution over the past five years has been tremendous. There's a lot of technology that came in our map making world, technology that has one aim only, that is to make our maps better, to make sure that we can deliver them faster, to make sure also that we deliver those maps at low cost. What you see here on the screen is not an animated map, it's key for my presentation to explain briefly what is this. This is probe data. It's the amount of probe data that we collect, in this case, it's Amsterdam. You see actually the map building up of Amsterdam.
This is played a bit faster than in reality. This is normally 24 hours. In 24 hours, you cover the map of Amsterdam multiple times. We are in an age of big data. Big data is hugely important for mapmakers like us. As you know, the applications that use maps have grown tremendously in number, in nature, and so on. We're not talking only about navigation applications. We have application in fleet management that's also pretty traditional. We have applications that are developed by developers that are using our Maps APIs, and those can be any kind of location applications that you can imagine. On top of that, we have new kinds of applications where the map is not used by the humans as such, but by robots. I'm talking about ADAS, advanced driver-assistance systems. I'm talking about autonomous driving, the self-driving car.
Now, all of that has made sure that there are a number of things that have changed. First of all, the quality of the map needs to be much higher than it used to be in the old days. Secondly, the time by which the map is updated needs to be much faster. We sometimes call that the cycle time. We want that cycle time to be as limited as possible in time so that you have always the freshest data that you can have. To do all that, you need an amount of resources, and because of the cycle time that is so limited, you need to have a very high level of automation so that your cost per modification in your database is pretty low.
I must say that we do that all quite well at TomTom, and that's one of the reasons why we, as an independent mapmaker, are one of the few companies that do that at scale. There's a couple of things which I would like to discuss with you, and that I think is worth remembering. First of all, mapmaking is complex. It's not that complex to us because we have the expertise, we have the technology to deal with it, but for any company in the world, it would be very difficult to produce what we already have in-house. It's not that difficult to make a map of Amsterdam.
If you imagine that there are 68 million kilometers of roads in our database, and then you imagine that you have to keep up with street names, with speed limits, with house numbers, with postal codes, with TMC codes, and I can go on for quite a while, and all those things need to be updated quickly and without that many mistakes. With the least mistakes possible. That makes map making at scale very difficult, and that's also the reason why we're one of the very few companies that do it. How do we do that? Well, we actually created a number of very specific technologies. In front of the building, you see one of those technologies. That is the MoMa van. That is the mobile mapping van. Actually, pull up TT 3.
It's very important to understand that that MoMa van didn't come out of nowhere, and it wasn't invented by Google, it was invented by us. In 1989, there was the first mobile mapping van, and it drove around in Amsterdam, in this city. In the meantime, I don't know which version we have today. It's version 50 or something like that. We perfected our MoMa vehicles, and they are very instrumental to capture images, to capture lidar data, which in turn is instrumental to make our HD maps and our navigation maps. That's one of the technologies we use. I will later on tell you more about artificial intelligence, machine learning, which we apply to automate our map making. I will also talk about Transactional Map Making system . It's quite a unique system that we have developed, and that was taken into production in the year 2016.
We developed before that for a number of years, and I think that's a unique system, and I will explain you what it is all about. Harold also already mentioned that we go beyond that, and we make actually sure that we create a map ecosystem with our customers. Our customers demand fresh data, high quality data, but they also imply that that data is used by their users, and their users give feedback. Because of that feedback, the map can be improved. We actually enable our customers to do that themselves, so that we create around TomTom a number of very important customers that actually contribute to our map database. Let me go back to the world of big data. Probe data is definitely one of the examples. More and more cars are equipped with sensors.
That sensor information is incredibly important for us to make our maps better and to do that faster and at lower cost. In that world of big data, I'll give you one example. I know you like numbers, right? Here are a couple of staggering numbers. We collect probe data from 600 million devices daily. Those devices include mobile phones, navigation systems, telematics systems, you name it. 600 million. If you want to have kind of an idea what that means in terms of how much data we get on a daily basis, we cover about 3.5 billion kilometers of road every day. I don't know whether it rings a bell, but it's about one time around the Earth per second.
That data is incoming and actually enables us to use that data to compare it with our map and to do all kinds of stuff. Here you see the probe data of a number of cities, and you see differences in the buildup. In New York, for example, we get the data about 50 times a day. 50 times enough data to cover the whole of New York. Singapore is even worse. It is more dense traffic. There it is about 600 times a day that we can actually cover the whole of Singapore per day. What we do with those data is we take that probe data, we compare it to our map, and we can find a number of things. We can find where there are new streets, we can find where the turn restrictions or the one-way restrictions have changed.
In this way, we can update the database. We can also compare the probe data of today with the probe data of yesterday to find out what has changed. We can really play ball very quickly with regard to updating our map database. Let me get back to the subject I promised to explain a little bit more, what is a transactional map making system and why is it unique? Most map makers in the past and most map makers today produce maps in a batched system. What that means is you take a map, you store it in a map database, you make all kinds of edits in that map database, and then once every month or once every week or once every quarter, you basically make a release of that map database.
You have stored all those edits, and in the release process, you quality check all the edits you have made. That is done in a batch process. There is fallout, you correct the fallout, and you go back to the release process, and at the end of the day, you release your map once a week or once a month or once every quarter. That's the traditional way of doing things in map making. The transactional map making system works differently. What happens in a transactional mapmaking system is that any change in reality can be taken through the whole process, including the quality checks, and immediately be available as an update of the map to the application. Which means that your database is actually continuously releasable. What that means is that with every change of your database, you create a new release, and that happens continuously.
You don't have to wait for a week or a month or a quarter. It happens continuously, and your applications can immediately use the freshest information that is in your database. That's quite a difficult system to make, and we were very successful doing so. As I said, we introduced it in 2016, and I will show you what the results of that introduction are and what that did to our database and to our productivity. Here's what we have today, and it is again, a staggering number, but we broke the record in August, with 2.35 billion modifications in our database in one month. That's an enormous amount of modifications in our database. 85% of those modifications were automated. 85% of those modifications went into the database in an automated way as a result of what we sometimes call fusion.
Fusion is actually nothing more than combining all different sources. That can be probe data, that can be aerial pictures, that can be images and lidar data from our MoMA van, and fusing those to create transactions that in itself have multiple modifications that are applied to the database. Now, in 2016, we were at the level of 250 million, roughly. We are now at the level of 2.35 billion, which is almost tenfold. What happened with the cost, because that's even more important, in the same period, we lowered the cost by a factor of 10. Which means that with the same amount of money compared to 2016, we can do 10 times more upgrades of our database, 10 times more modifications in our database.
Which is kind of a fantastic result, and the foundation of that is our transactional system and a much higher level of automation driven by the usage of machine learning. How do we use artificial intelligence? Well, I can give you plenty of examples, but I'll give two examples which are highly relevant. In the HD mapmaking world, but also in the ADAS world, it's very important that we capture the signposts. The signposts are captured by our MoMa vehicles. They drive, they have lidar information, and they have visual video information. That lidar information and video information is checked for reflectivity. As you know, the signposts reflect a lot. We identify those signposts fully automated, and we can actually recognize whether it is a speed restriction, whether it's one-way restriction, whatever.
We have a number of signposts that are in the catalog, which are by machine learning detected in our database. That's one example. Another example is through the probe data, we will know, for example, where there is a new road. We need to pick up the geometry of that road. Well, the way we do that is we use our satellite images, and with machine learning, we can actually pinpoint first, based on the probe data, where that new road probably starts, and then the new road is fully automatically created in terms of geometry. Those are two very simple examples of how we make a higher level of automation in our mapmaking world possible. Let me come back to the map editing partnership.
What that does, all the previous things I talked about, what that does to our database is indeed that we have a higher quality database that we can deliver very fast to our customers, which integrate that database in their applications. Their applications are used, in some cases, by thousands of people. Take Uber as an example. Uber is using our maps, and there are many drivers that basically know whether there is a mistake in the map or see whether there is a mistake in the map. Traditionally, what happened is we got that feedback through Uber towards us. We put it in our production process, and then basically we provided them with an update. The way it works today with those map editing partnerships is that they don't have to return us all that information.
We actually have given them the tools and the training to, on their own, be able to correct immediately the database. You can imagine what that does. It increases the quality, and again, it lowers the cycle time. It makes sure that we have better maps faster. That's what we do with the editing partnerships. It gives shorter cycle times, and it makes the map better, which is crucial for all the customers that we have. They are more than glad to be able to increase the quality of the database and reduce that cycle time. I think as a conclusion, I hope I've given you a glimpse of all the technology we are using and the processes we are using with one goal in mind. To make maps better.
Higher quality, which in robotic world is even more important. Willem will talk about the HD map in more detail. Also for navigation and other applications, the quality is increasingly important. The cycle time reduction, which means that you can have much faster updates, that you can incrementally update the applications, is crucial in an online world. As Harold talked about, we are more and more going in the direction of online applications only. That means that the expectation of the end user is that your map is just continuously up to date. With our transactional system, we can make that a reality.
It's not true for all the attributes we have currently in the database, but as you know from our traffic, for example, there we already have online information available, and we can do the same thing for many attributes and content in our database. At the end, we do that all for a much lower cost per transaction and per modification than in the past.
Thank you very much. It's François-Xavier Bouvignies from UBS. A couple of questions. You talk about your backlog that's going to be your new KPI, removing the order intake. At the same time, you say that you expect overall growth to accelerate beyond 2021, which would imply overall order intake to increase in the coming couple of years. My question is, what is the main driver? If it's the case, is it the case, first? If it's the case, what is the main driver that would increase your order intake in the next few years? The other question is, if your backlog is EUR 1.6 billion and you have take rate assumptions, you have the volume assumptions, what is the take rate assumptions you have today for 2021 given the visibility that you have? That's the first question.
Willem take first, then Stefan.
We'll discuss take rates in detail later today. The take rates now are what we have for the contracts that we have signed, take rates are anywhere between 30% and 40%. That depends on brand and car type. We foresee take rates to go towards 80% in the next decade. That's not the explanation of the backlog, because the backlog, we see today's take rates. That, again, that can be low 30% or it can be high 30%. The acceleration of growth, one of the drivers why we expect acceleration of growth is take rates. Antoine, our Head of Automotive, and Kees, our Chief Product Officer, will talk about that.
Yeah. I think the other thing that you see in that forecast is over time take-up of technologies for automated driving and ADAS. That will really start to kick in the midterm. That's not for this year or next year, but slowly but surely, that will become mainstream application for most vehicles.
One of the things that we will see is that regulation is also helping in our favor. In a couple of years, the EU wants to make it mandatory for European cars to see speed limits. That can be done by camera, but that can also be done by a map. Especially for the less expensive cars, the latter will be more the default solution. That could also increase the take rates. At different price points, obviously, but that will generate new income streams.
Two quick follow-ups. Hello?
You ran out of batteries.
Yeah. Hello. Yes?
Yeah.
Perfect.
Yeah.
Two quick follow-ups.
Keep going.
On your market share, obviously Android, Google made a lot of noise last year with Renault. Given that you have more long-term assumptions now, do you expect any change of the market share in the next few years to justify this 15%? Is it based on stable market share? Do you expect some disruption there? The last one, all of you talked a lot about cost reduction in your map-making platform. I just wanted to know, how is it translating in your operating expenses? How should we think about the profitability of TomTom in the next few years? We see that sourcing is going down in terms of cost processing as well, but engineering seems to offset all these benefits. I just wanted to double-check with you. Thank you.
I think we're fairly disciplined in how we run our cost. We are at that tipping point where we are generating serious cash this year for the first time. Longer term, I think we will continue to generate cash. We see it also a bit as a boundary factor that we want to run a show where we have positive free cash flow. I think we can do that. I think that's built in our D&A, and that's how we see the business going forward. Those map-adding partnerships are important; they're not in itself reducing our cost base. They're making for better maps and better customer service, that's a selling point in its own right. You don't see any meaningful, at this stage, because of the volumes we're having, you don't see a meaningful reduction in the cost base.
It's really giving our customers the ability to inject data into the database that matters to them. That is an important capability that we can give to our customers. As a result, the map will get better, and that will raise the appeal for doing business with TomTom as a location technology provider. You will see a very firm transition from manual labor to automated processes, and for that we need engineering. In the HD mapping space, you also need engineering to do that cost-effectively and to a high degree of automation.
On the market share?
Yeah. Let me touch on the market share as well. What Harold explained in his presentation as well is that for traffic specifically, we have 80% market share in Europe. I don't think there is a lot of room to grow that further. There is room to grow our market share in the U.S. That will happen based on the order intake. We've signed, let's say, two of the three major American OEMs. That will increase our market share for traffic. On the mapping side, today, HERE is the incumbent. We have been gaining market share, and we will continue to gain market share. We are the follower there. On software specifically, software today is a bit of a generic term, let's say the navigation suite or IVI, that is a very scattered area.
No one has more than 15% market share. We have a double-digit market share. We believe, and again, Antoine and Kees will talk about that, we believe we can significantly grow our market share in the application layer and what we can do there.
Excellent.
Yep.
Yeah.
Yeah.
Marc Hesselink, ING. I would like to come back on that market share. Obviously, a lot is changing in the competitive environment, and Google now coming into the picture, maybe if everything that's being announced, if they would move everything to Google, maybe they have 30% market share. Is that your assumption in the long run, that there will be those three players, yourself, Google, and HERE, and all three of them taking that 1/3 of the market? Is that your assumption behind the 50% CAGR in the long run?
Well, in our CAGR, we don't anticipate a massive increase of our market share. It's not that we think by 2030 we have 90% market share. That's not what you see in those pictures. What you see in those pictures is a stable, perhaps slowly growing market share, but much broader applications and a higher attachment rate, and new applications coming in. Specifically for ADAS, HD, but also IVI, so a more integrated approach to the user interface in the vehicle. Kees van Dok, our Chief Product Officer, will talk about that, and we see great opportunities there as well in gaining market share. If you do that, then you have a much broader part of the whole software stack in a vehicle.
Makes a huge difference, of course, if you only provide traffic or if you manage the whole stack, especially when that goes online, which is what our expectation is.
It doesn't also imply that, because the other one could also happen, like Renault, a big client of yours, moving away to Google? That could actually imply that over time, then your market share on-
Yeah.
goes down.
We need to see how that will pan out. We need to see how all that will pan out. We are obviously aware of what Google is doing. They have some early success with some OEMs. We don't think that the whole market will go to one player. That's just simply not how it's going to work. The industry will always want alternative solution and alternative vendors. We believe we are very well-placed to be that vendor. There's a number of drivers there. First of all, we see higher degrees of integration in the vehicle, all sorts of silver boxes that need to talk to a unified user interface. That is very difficult to do that with an application that's basically taken from a mobile phone and put in a dashboard. The integration is important. Electrification plays an important role.
That's from a product perspective. Also from a business perspective, we just do not think that all car makers will hand over the keys of their user interface, an important part of real estate, to a company that sooner or later will start competing with them, both in self-driving or use data for alternative business models. That is not a healthy situation, and we don't think that it will happen. At the same time, of course, I think there's good news as well. We have been arguing for a long time with our customers that we need to change the way we're working, that the applications that are currently in the market don't live up to end-user expectations.
You see that a lot of customers who paid for a navigation system or an infotainment system in the vehicle are actually not using it, but are referring to a mobile phone that's more trusted. Well, that is not good. We need to fix that. Our customers understand that they need to fix that, and there's a big drive to going to an online model now where you can deliver an end-user experience that is compelling and that lives up to expectations of also a younger audience, a more modern audience, and so on and so forth. We think there's a great opportunity now to step in that void and offer a great alternative to a product from a vendor that is not fully trusted.
Okay, thanks. Maybe a bit more on the short term, you explained the part on the automotive and the location technology category, 10%, but there's also big moving parts always in the deferred revenue. How will that be in the next coming years? Obviously the impact therefore on your free cash flow in the coming years.
I have first a related question to that.
Yeah.
You're good at that.
No.
No. I'm about maps.
You do maps.
I do maps.
You do the I first.
Thank you. For this year, what we've said is that we expect roughly a net addition of EUR 60 million, EUR 100 million being deferred in automotive, EUR 25 released in the consumer, and EUR 15 million in enterprise. For next year, that number will be a bit different. Automotive, again, a net defer, maybe not EUR 100, but EUR 95. Consumer will also go from EUR 25 to EUR 15 because the install base is declining, and also the term of deferring and releasing has shortened. We'll go more to EUR 15. Again, in enterprise, we see something like EUR 25. These trends will continue. I think that for automotive, the net amount that we put on the balance sheet will decline. Where we see EUR 100, we're going to EUR 95, and it will soon after that further decline.
The net release of consumer will decline as well, 15 towards 10 and going forward. Enterprise is a bit different beast because that's more related to large customer additions, et cetera. After 2021, I don't expect huge differences, et cetera. With the knowledge that we have today, it will be driven by automotive and consumer. Anywhere like EUR 60 million-EUR 70 million from 2021 onwards.
Okay, thanks. The final question is on the You gave the first step towards maybe cash returns in the next years via buyback. What are the things behind it? Is it then something, if I read next year, is that something that you discuss at the full year update, the full year 2019 update, then probably starting in 2020? What I mean with what's behind it, is there a sort of minimum cash level that you would like to have on your balance sheet? Is it dependent on the free cash flow that you're going to generate in the next coming years? What's behind your thinking there?
Yeah, it's more an April thing, with our AGM. I want to push that out a number of two months. It is what it is. It is the investments that we need to make towards It's the mapping and the acceleration that we see there, or the pace of investment together with the free cash flow that we see. It's a combination. It's not a hard science or what have you. We feel that four months ago, we did what we had to do. We gave back EUR 750. If we go over the half a billion EUR of cash on our balance sheet, that is kind of a moment to reflect again. Again, with the AGM, we will give an update on that.
We know more about the whole business and thought about it a bit better, and I think it's a good moment. There's no right number for how much cash you want to keep on your balance sheet. I think it's the right number at the moment. As you know, there's a lot of turmoil, there's a lot of change in the industry, great opportunities, but also that brings with it uncertainty. We also want to be able to react if that's needed, and we don't want to be forced to do dumb things because we lack the firepower, or we can't do what we want, not ready to grab market opportunities. I think around that number, you're kind of okay. Again, I also said we want to be disciplined. We want to generate cash. I think we're well on our way.
Our longer-term plan shows us also the ability to keep generating cash. We feel good about that. We're strong. We can move as needed. If the market opportunities open up, then we have the firepower to do things. There's nothing on the agenda, but just to have that flexibility, I think, for this type of business is good.
Okay. Thank you.
Now let-
Wim Gille, ABN. Two questions. First of all, IFRS isn't doing you guys a lot of favor in terms of revenue recognition, but also the IFRS definitions of backlog can be quite harsh on you guys. The backlog definition that you use, is that a pure IFRS definition or is it that you have constructed a TomTom version, basically adding up all the IFRS or expected IFRS revenues and come to that number? The reason why I'm asking is because, obviously, depending on the contract structure, IFRS has a tendency not to account for stuff which is actually going to be there. That is my first question, probably for Taco. I've got a question.
Thank you.
for Alain.
Yes.
You guys made a great progress in the transactional mapmaking business. Do you guys have any feeling on how you stack up vis-a-vis to competition? Where is HERE going on this particular topic? Where is Google on this particular topic? Do you guys have a view there, how far advanced you are? Also how difficult is this to replicate for a number of startups, which are basically just coming around the corner and some of them are quite well-funded. Do they have kind of a fighting chance in getting to the position that you guys already have?
You want to take this?
Yeah. First of all, what we have in our annual report, the backlog is something different. It's a combination of enterprise and automotive and does not include all the contracts, because it's only the IFRS 15 contracts that are in there. It's a very relevant KPI for the regulator, but not for the analysts and the investors, because you can't do anything with it. What we decided is to focus on automotive only and just look at the awards and the contracts, and based on the assumptions made on the tenor and the car sales and the take rates, and apply the committed prices. If you look at the total order intake, operational IFRS will be the same, but only the spread over the years. Obviously, in IFRS, it's different, so less early and more at the end. Does that answer your question or?
Yes, I think on top of my mind, you have a backlog of around EUR 1 billion or so in your annual report. That's kind of a completely different number, different definitions.
Yeah.
We should continue to ignore that number.
Yeah. Please do.
Thank you very much.
With regard to the mapmaking platform, we think we have a pretty unique platform. Let's talk about Google. I don't know what Google has in the kitchen. It would be wrong to say that they don't have a transactional system. I don't know that, right? What I do know is that HERE does not. What I also know is that you absolutely need a transactional system if you want to have a future. As the world is moving online, cycle times and updating information are so critical that you can't get away with batch systems in the future. How much time it would take them to develop it? I can tell you that it took us four years with a lot of people. It was a big investment. Maybe they're more clever, maybe not.
With regard to your question on startups, most of the startups in the mapping world are one-trick ponies. They have one specific area of mapmaking, and they demonstrate that they're really good at that. What we're talking about in the environment of Google, HERE, and ourselves is mapmaking at scale, which is a completely different game. Before a startup, and if you say well-funded, most of those startups are funded with EUR 90, EUR 100 million, maybe EUR 200 million. That's by far not enough to get where we already are. Actually, a proof point of that, there's two proof points of that. If it would be easy for a startup to do what we do, you would have plenty of them that already are somewhere in competition with us, and they're not. That's one proof point.
Another proof point that is even more impressive, at a certain moment, Google decided to make their own maps. Before that, they were using Tele Atlas maps or TomTom maps. It took them seven years to achieve what we have. Don't forget, it's 68 million kilometers of road that are kept up to date. Small, little startup. Never say never. We also started as a startup many years ago. Is it possible? Yes, it's possible. We're not talking about EUR 100 million. We're talking about way more before they are where we are. In the meantime, as you have seen from the graphs, we're moving really fast. I think it's not very likely in the foreseeable future we would see that coming.
Okay. There. You're allowed to throw with that.
Thank you. Andrew Gardner from Barclays. Just to follow up on that question and then some of the points you were making earlier, I think both you, Harold, and Taco, on market share. If we go back a few years, you were willing to talk about you having 20% market share of maps in car and HERE at about 80%. Taco, you didn't really want to touch on a number there. It feels like it's been moving in your favor, but perhaps not dramatically so, given how long the development cycles, the product cycles take in the automotive space. I'm just wondering why, given what Alain just described in terms of the strength of the transactional database, and particularly relative to HERE, who is seemingly the primary legacy competitor, why are you not more optimistic in terms of the potential for market share gain in automotive?
We are. I think the leading indicator really on the West Coast, they are faster, they have a deeper understanding of technology and how that goes. Compared to the West Coast company, I think the car makers are typically a bit behind, a bit slower, and for all the right reasons. That's not a criticism, that's just the nature of that business. It's a different business. You want to be a fast follower. You don't want to be an innovator, probably, if you are a car maker. If you look at what we have achieved on the West Coast, we're very proud that we are now really embedded with the leading technology companies there and entertain very good relationships. That's a mutual benefit. We learn from that. We get faster, we have good visibility of what's happening.
I think that will translate also, and it is translating already, in our firepower in the automotive industry. That will trickle through, but it takes more time. In the last couple of years, we've seen market growth, market share growth in the automotive space with some very significant wins that have always been customers of our nearest competitor and decided to shift their business for a large proportion of their business, it's not always everything, but at least a large proportion of their business to TomTom. I think that's a good sign. We're confident that we will continue on that path. We're not putting in, as I said earlier, 40%, 50% market share in our longer-term plan.
There's just simply not enough evidence to say that that's going to happen, and we want to be cautious, and we want to be careful, and it's certainly not how we are planning our expenditure. I think we are well-placed to take advantage. We have a lot of cooking in the kitchen, really cool stuff, really excited about that. We list a little bit of that later today. You get a bit of a feel how the thinking is developing. We want to really reach out, not just to the car makers who are paying the bills at the end of the day, but over the head of the car maker, also the end user, and make sure that we can supply our customers in the automotive space with an end-user application that really cracks it.
That people actually use, and that they find charming and entertaining and fast and of great quality. I think that's a challenge that we have defined for ourselves to be able, in that position, to help our customers to keep control over their dashboard with end-user applications that really matter and that are first class and can compare favorably what you have on your smartphone, whether it's from Google or Apple or any other vendor. Our target is to put something in that screen that is as good, if not better, and better suited for use in a car environment than a application that's typically developed for a mobile phone application. That's the longer-term vision. If that sticks, if we can get that through, then we have all the underlying technologies, the vision, the experience with end users to be able to play that role.
I would find that very exciting if we can get there together with our customers to start delivering products that are really top-notch, that are charming, delightful, and do what customers are expecting them to do. That's a great challenge, and I think we've got what it takes to take that on. I think that our customers see us also in that respect, as an alternative to taking a full package from someone else that they have limited influence on customization, product roadmap, and so on and so forth, with a supplier that could potentially sooner or later start competing with them. I think that's a great product proposition and a great business proposition. Now we need to get on in convincing and working with our customers to make that vision a reality.
In the meantime, of course, we have a lot of stuff already now ready to ship, ready to move. That's our current business and that will keep going. There's a whole new class of products and user experience that we want to bring to the automotive industry. There's a whole new thing that we want to bring to our enterprise customers with a more collaborative way of making top-notch maps. Again, with a view to circle those customers who have, for one reason or the other, do not want to commit to a platform of one of their competitors. I think in that space, in that sweet spot, there's a great role for us to play. We have what it takes in terms of heritage, culture, output looking, customer focused development. Excuse me.
I think that's where we want to hone our skills and want to progress in the years to come. I think that will work. If you look then at the outlook for the next 10 years, you see a very significant number that's partly coming from that integrated product and dashboard, partly coming from new products like ADAS and HD, where I think we're on the right track as well, and also partly coming from, of course, what we think we can achieve in the enterprise world.
Just a quick follow-up. You mentioned the success with the West Coast companies that's giving you. You also highlighted Apple on one of your slides. I'll ask the question, I'm not sure whether you can answer it, but they themselves are also doing some map development.
Yeah.
What type of visibility can you give us into the strength of your relationship there and longevity, given that potential conflict?
Well, the only thing I can say is Apple has been a fantastic partner for us. Absolutely, we've done a long-term deal that will run for years still. They kept their part of the bargain, we kept our part of the bargain. There's a lot of trust and collaboration going forward. There is an end date to that contract, of course, I am committed and hopeful that there will be something in the future with Apple as well. I can't guarantee that, I think if I read the tea leaves correctly, I think there will be a future after the expiration of the current contract.
It will have a different nature, a different role we will play, I think there is hope and opportunity for a new type of relationship, very much along the lines what I just said, being in the middle of that ecosystem, where everybody kind of looking for a good alternative, an alternative map that is of high quality, but where they don't have the burden of collecting the data, do the quality control, and so on and so forth. That's a space that I think we can occupy. That will take some time, but it's clearly part of the strategy going forward.
I'm Anders Troelsen. I'm heading up our business unit, Enterprise. I'm based out of San Jose, at the heart of Silicon Valley, and over the next 15 minutes, go through with you what we're up to in the enterprise business unit space. In our business unit Enterprise, we work with a lot of different type of companies, and we are extremely pleased and proud that so many companies have selected TomTom as their trusted location partner, whatever that is, for fleet or logistic, web or mobile applications, on-demand or ride sharing applications. What a lot of people are not aware of, if you put all that usage together, it basically means that over 1 billion people can access our products every single day through all these different type of partnerships we have created over the years. That, of course, in itself is pretty amazing.
I want to go through with you two, you can say high-level go-to-market products areas. The first one is uncompiled maps and traffic data, and then later I will go through our other product offering, the Maps API space. Let's start with uncompiled maps and traffic data. I think the best way of looking at that and explaining that is you look at our maps in different layers. Whatever that base map, street map, street names, point of interest, like hotels, restaurants, traffic or navigation, all that comes together. As an application developer, you don't have to use all the different type of layers. Let's take an example. If you want to create a store locator, you probably wanted to take the base map, the whole street network. You probably wanted to take the street names.
You definitely want to take your own store from a point of interest, and you put that together as your application for a store locator. Let me give you another example, and in this example, I will use Uber. Uber is taking all the layers we have in our uncompiled maps and traffic. We put that into the phone. Uber have decided that they wanted to create kind of how their own map looks in terms of colors and how it behaves. They've also created, you could say, some interface between the driver and the rider to have some kind of interaction, also get some stickiness for the driver to use the Uber application. We put that in as well. Basically what comes out of it. Oh, sorry. What comes out of it is. Oh, not working.
What comes out of it is the Uber driver application, which is being used by millions of drivers every single day. When we ask our customers what is really important, that is map freshness. We can tell our customers that we are doing 2.4 billion modifications every month, is absolutely helping our story towards our customers. Of course, there's a lot of other things which is important, as you can see in the slide, in terms of different sources, which is coming into our map database. As you heard previously also from Alain, that we have all these type of partnerships, which basically means you have access to 600 million live devices, which is providing us with probes every single day. That in itself is a very unique proposition to the market and to our customers, and is actually helping us a lot winning deals out there.
We didn't stop there. We're kind of sitting down and say, "Hey, how can we help our customers helping their customers even better than what we're doing today?" Basically trying to delight, you could say, through the whole channel. That's basically how we came up with what we call, and you heard before, the map editing partnership. We've basically been taking our tools, giving them to our customers, our strategic partners, and letting them edit directly into our map database. Of course, the whole idea behind it is that we can turn these edits around much faster, and we can have a much fresher product. Of course, our partners are very interested in having that freshness so they can provide the best product to their consumers, and again, in the market they are in.
What you see here is the tools, just an example of you changing a street name, which you can do directly into the TomTom map database with using the TomTom tools. We have trained our partners as they were exactly TomTom employees. All the quality rules, all these type of things, of course, is in place. In the beginning of the slide, you saw a lot of different type of logos. As you heard also me talking about in the beginning, like some of the segments we are active in like fleet and logistic, web and mobile, and analytics. Let me run through with you a couple of example how our maps are being used, with different type of our partners. The first one I want to show is PTV, based in Karlsruhe.
PTV is very strong in what we call traffic management solutions. Of course, what you see here in the picture is the traffic center, and of course, maps, traffic is extremely important for all these type of traffic management solutions. Again, they're using 100% TomTom for all these different type of implementations. Another example I want to show you is Apple. Apple and TomTom have been having strategic partnership for many years, and for many years to come. Also if you see here, if you take Apple Maps and you open it, you click in the upper right corner on the information, and you go to the next screen, you will see a TomTom logo. Also Apple is very proud of showing that I'm working with TomTom and showing that to all of the users of Apple Maps. The next one is SAP.
SAP has been taking all our base map and addressing and created a global geocoder, which both can be used by SAP themself internally, but they're also exposing that global geocoder to SAP's users, and then you can basically get access to it and pay via SAP if you want to use that global geocoder. The next one is Pitney Bowes, an American company. They're very strong as well in Europe under the brand of MapInfo, which is being used a lot with different type of municipalities for road planning and these type of tools. They're also very strong in the insurance space. Basically what they're doing, they're taking all the map and traffic data from TomTom, and then they're conflating on top. Again, if you think about the layers, they're basically putting their own layers on top of different type of information.
Let's take an example. You're an insurance company, you want to put a price to a business or real estate or your own house. Basically combining even with flooding, crime in the area, how far are you from police station or fire station, they conflate all that data together and bring that solution to the insurance companies. I also have a small video there I want to play, basically explaining how we're working closely together with Pitney Bowes and why that relationship is important for Pitney.
The Pitney Bowes mission is to organize and manage global address data and to be able to then provide attributes and enrichment data around those addresses.
Pitney Bowes has made the decision to do business with TomTom a number of times over the last 20 years. The decision came down to very complementary business models. TomTom enables Pitney Bowes to execute our strategy because of the investment that TomTom makes in building and maintaining global maps.
We work in a number of key verticals. Our primary one is insurance. Our insurance clients use address validation and cleaning, being able to manage so they have a single view of that customer. Even more importantly is our geocoding and our location data. TomTom collect really, really valuable data around addresses, around streets, around points of interest. The fact that it's global in coverage, and it has a consistent data model is really, really important to that relationship.
At Pitney Bowes, one of the most exciting things as we look forward is working together with TomTom to build out a complete and current, highly accurate and precise global addressing data set. I really think the most exciting time in our relationship is ahead of us.
Okay, that was kind of a little bit about our product offering in the uncompiled maps and traffic area. As I mentioned in the beginning, I'll go through our offering for the Maps APIs. Let's start asking the question, what is an API?
Is the application programming interface or API. It's the engine under the hood and is behind the scenes that we take for granted. It's what makes possible all the interactivity we've come to expect and rely upon. Exactly what is an API? It's a question everyone asks. Okay, not really, but we're glad you did. The textbook definition goes something like this. In computer programming, an application programming interface, API, is a set of routines, protocols, and tools for building software. An API exposes a software component in terms of operations, inputs, outputs, and underlying types. Okay, to speak plainly, an API is the messenger that takes requests and tells a system what you want to do, and then returns the response back to you. To give you a familiar example, think of an API as a waiter in a restaurant.
Imagine you're sitting at the table with a menu of choices to order from, and the kitchen is the part of the system which will prepare your order. What's missing is the critical link to communicate your order to the kitchen and deliver your food back to your table. That's where the waiter or API comes in. The waiter is the messenger that takes your request or order and tells the system, in this case, the kitchen, what to do, and then delivers the response back to you, in this case, food. Now that we've whetted your appetite, let's apply this to a real API example.
Yeah. What is a good example of real examples? Look at it like that. We have the kitchen, and basically, we now take the uncompiled maps and put it into the kitchen. We then have hired the best chefs in the world, which of course, is our engineers, and they're doing all that cooking for you. Basically, what it means for you as developers, it means much faster time to market using the APIs versus taking and using the uncompiled maps. Of course, the unique thing for TomTom is we have both, right? We are both offering for the market, whatever you think, whatever company you are, whatever application you want to build or solution you want to set up, we have both offering for you. Again, with the APIs, it's a very easy way to get to market very quickly.
We are estimating that the overall or the total market of Maps APIs is around EUR 1 billion, and it's growing as well. We started recently as TomTom with our Maps APIs, which basically means that our market share is tiny today. We definitely have high expectation in the overall Maps API space. If you look at our product portfolio on the Maps API, we have search, but of course on the search you have different type of functionalities like geocoding and reverse geocoding. Routing, different type of functionalities as well. I'll come back in a second and show an example of EV routing, which we released recently. You have map tiles, you have the traffic, and in the end you have the Maps SDK. Let's take again the example I used for the uncompiled maps for store locator.
In this case, you just hit the search API for your store and you hit the map tiles and basically already there you have your store locator. Again, hopefully, that's a good example of showing that with Maps APIs, it's very simple and very easy to use and fast go to market. Let's go into the routing and show you an example for the EV routing. Here we're calculating a route from our office in San Jose to Los Angeles. You're driving a BMW i3, and the blue polygon you're seeing immediately is the range you're having for your EV car. You can see that, of course, that when you're getting to the range, outer range of the blue polygon, then you have to find or search for an EV station.
That we're also providing part of our EV offering so you can search for the EV stations. We will tell you if the EV station is open. We will also tell you if the EV station is free so you can get there and start charging your EV vehicle. With that, of course, we're helping all the EV drivers getting from A to B in a very smooth and easy way. If you look at our Maps API channels, we have high-level 2 channels how we go to market with our Maps APIs. We have our own developer portal, and then we have our enterprise unit for the more, you could say, larger strategic deals for going out to the market. If you start with the developer portal, you just go to developer.tomtom.com and you get access to our developer portal.
Look at it a little bit like a store where you can go and browse around and figure out what you want to buy. You can go in, sign up, test, and play with all our APIs. The business model we have introduced here is what we call pay as you grow. You start free, but when you start growing your business, you start paying for the API and the usage. Again, it's super simple. You can just put in your credit card. You don't have to talk with us. You can do anything on your own. If you want to talk with us, we like that as well. You can, of course, contact us and we can do that online and help you build your application. Again, very easy and very simple. The other areas, as I mentioned, is the enterprise area.
One specific contract I want to highlight is our contract with Microsoft. We started our contract with Microsoft a while ago where we, as TomTom, announced we are putting our Maps APIs in Microsoft's cloud, so Azure, which basically means that all the Microsoft internal developer have access to the APIs. We also agreed with Microsoft that they could bring it to the market under their own brand. They bring our Maps APIs under the Azure Maps brand. Microsoft is then adding some other interesting things in the product offering themself. We recently also announced between TomTom and Microsoft that the Bing Maps and Cortana will also move to TomTom. I think it's fair to say that our relationship with Microsoft is very good and very solid, and wherever Microsoft have a need for any type of location, maps, traffic, they're using TomTom.
Instead of I'm continuing explaining what Microsoft is doing, it's much more interesting to hear it from Microsoft themselves. I would like to invite Chris Pendleton, Head of Azure Maps, to stage.
Thank you. I want to thank TomTom for inviting me to the stage. This has been a very lucrative partnership for Microsoft. I've been in the map space for 20 years now, varying degrees of impact across the company. I've been at Microsoft for 17 years, three years at a company prior to that. I can fully appreciate Alain's presentation. The ability to make maps is hard in and of itself, but the ability to make them fast and keep them fresh, it's exorbitant in terms of resources, cost, and the amount of data required to actually keep them up to date. A few years back, just to give a little bit of history, a few years back, I ran the Bing Maps data ingestion team. I was responsible for keeping Bing Maps' data up to date.
We would take quarterly drops from our provider, and it would take us anywhere from 6 to 9 months to get that data out. If you do the math, you get a drop every 3 months, then it takes 6 to 9 months to get it out, so we miss one or two drops in the process. It didn't work for us. The product of Azure Maps was created effectively for two reasons. The first of which was to bring location natively to the Azure cloud. In our conversations at Microsoft in terms of competition, we talk about the three clouds that are currently in competition with one another. Azure Maps is natively integrated as a first-party product of the Azure cloud. When we talk about Azure and all of the Azure wins, we're talking about location through Azure Maps.
Okay? As you start to see these wins and announcements coming through Azure that include location and effectively Azure Maps, we are effectively talking about TomTom. Right? As Anders mentioned, we struck this lucrative partnership that brings TomTom's APIs to the Azure cloud, and we wrapped those up and made them a part of our platform. This is the definition that I wrote for our documentation. It's super nerdy. Basically, Azure Maps is a collection of location technologies for Azure customers. As Azure grows, our location capabilities need to grow and keep up with our customer needs. Okay? We want to make that simple for our Azure customers to use. The way that Azure works is you actually subscribe to an amount, a pre-committed amount of Azure, and then you start spending against that commitment.
When companies start to look around over the one, two, three, five-year commitments and they see Maps there, it's a pretty easy decision because they've already pre-committed for amounts that they need to spend. They just go to our portal, and they start using Azure Maps. In fact, somebody on my team is dedicated to every morning looking up and seeing who our new customers are because every day we actually don't know until we look up the report. There's no negotiating for us. We don't get pinned against Google Maps anymore. It's an Azure discussion. It's higher level. Azure Maps is the native location platform for Azure and the Azure ecosystem. We actually sit inside of Azure IoT organizationally, and within Azure IoT, there's a group called Azure IoT Mobility.
Azure IoT Mobility is made up of the Microsoft Connected Vehicle Platform and Azure Maps. Okay? That gets me to the second part of why Azure Maps was created, and that was to power a direct vertical integration into the automotive space for the Microsoft Connected Vehicle Platform. Last week, two weeks ago, 10 days ago, we announced that TomTom's native integration with their navigation kit would be a part of the Microsoft Connected Vehicle Platform. This is a significant move, right? The Connected Vehicle Platform brings edge computing into the vehicle. Okay? Edge computing means it runs in the car. It can actually run AI in the vehicle. It can make decisions in the car. Right? It's trained by cloud computing. Get data, put it in the cloud, you train it, you install modules in the car, and now things are happening in the car.
We call that the automotive edge. Okay? Azure Maps sits quite nicely right in that ecosystem as well. It is an ecosystem play. We are inviting other partners to participate in the Microsoft Connected Vehicle Platform, TomTom being one of the most prominent. Okay? Azure Maps is a horizontal set of location APIs or services for Azure customers and a vertical integration for our Connected Vehicle Platform inside of Azure IoT. I decided to include the customer presentation deck that we give at executive briefings. We have an executive briefing center at Microsoft. Every day, dozens of executives fly from around the world to Microsoft, and we treat them to a day of the products they want to hear about. This is a significant portion of that deck. It's really long and technical.
I decided to focus on some of the more important facets. I mentioned we're in Azure IoT. I've been speaking about this location of things concept for the last year or so, and it's effectively IoT plus location. IoT has this proliferation of devices, and when we talk about IoT inside of Azure, we talk about these 3 stages where you've got device sensors that are the actual devices that are sensing things. That data comes up into the cloud, and with that data, you can generate insights. Okay? You've got presence sensors in rooms, so you can know that certain rooms have people in them. You've got thermostat systems in houses. You've got thermometers out in fields.
There's a slew of Azure IoT scenarios that generate a number of data insights. What we do with those data insights, or our customers are doing with those data insights, are then making data-driven decisions. Okay? If we do this fast enough, there's a stage 4, That is predictive. Okay? Once we have enough data and we've generated enough AI, we can start to predict what's going to happen based on machine learning. The location of things gives you that important piece of information as where is it happening, right? I can tell you that there's a field out in Eastern California that is too dry based on the thermostat and the weather patterns. The sprinkler system needs to turn on. That's just one piece, right? Where is that sprinkler? Which one is it? What field does it cover?
That's where the geospatial facets come in. When we talk about mobility, we talk about fleet management, managing where trucks are at any given time. These are IoT mobility scenarios and how they come to life. This is an actuation, so I think this is as technical as I'm going to get for you. The idea here is that if you take telemetry from a vehicle, and of course, that means the vehicle is connected to the Internet, and the Internet can publish that data to the cloud. You publish those messages up through IoT Hub. IoT Hub is one of the core products for Azure IoT. It basically can receive messages from devices. Okay? From IoT Hub, we have a listener on Event Grid for changes. In this particular scenario, what we're looking at is a geofence.
Geofence is an invisible fence around a particular area. We've just fenced off this particular area so that when the truck leaves the fence, we get a notification. It sounds simple, it's actually built in geography. It's built in mathematics. We need to know where the truck is at any given time. We need a pulse for where it's going, where it is, when it's there. We get a pulse reading. As soon as it is no longer inside that polygon, an alert kicks down through an Azure function, kicks off alerts through Event Grid. These alerts could be text messages or a sequence of events after that in programming languages that can then do other things, notify the driver that he's off route, whatever other things you want to cascade after this.
Azure Maps brings the canvas, which is powered by TomTom, and it brings the geofence, which is the geospatial representation of that invisible area. The data gets dropped in Blob storage. Blob storage basically means I'm going to use that later, probably for training purposes. Okay. I want to train some ML, machine learning later, create some AI. I have an iconic slide just like Anders did. He mentioned we've done some things as well. When I present this slide, we've got a robust offering for Azure Maps and Azure customers, right. We've got maps and satellite imagery. The maps are powered by TomTom. Every map TomTom has, I have coming through Azure Maps. The SDKs. Our SDKs we actually built based on open source. We took web SDKs and rolled them, and we're pulling in TomTom services through the web SDK.
As Azure customers are taking Azure Maps SDKs, they are pulling through TomTom services. Routing. Anders covered this. Didn't do it justice. There is so much complexity in the routing API that TomTom provides. Shortest and fastest is sort of the simple ones, but also route optimization. Give me a bunch of points, tell me the order in which I should go through them. The isochrones, which is how far can I get in one minute from here, in five minutes from here, in all directions? Ends up creating a polygon. Electrical vehicle routing. All of the turn restrictions are included. There is a multitude of scenarios. In fact, I have a huge slide that has all of the routing capabilities, taxis, HOV lanes, vans, bike routing. All of this is included in the routing API powered by TomTom.
The search API is everything that TomTom has in their data corpus, I could search for it. Every address, every point of interest, business listing, every landmark available to Azure customers. The spatial operations is a unique thing to Microsoft. We actually built this, and it includes our geofencing service. We're rounding out the portfolio of offerings from TomTom with some additional capabilities. This gives us spatial analytics so that we can plug in different data sets and do analytics based on spatial information. Anytime geography comes into play, we use TomTom. The traffic data, best of breed across the board, powered by TomTom. We get this as flow and incident data. We also get some measurements. If you're approaching the back of the line for any port of congestion, it'll give you a measurement distance and time from where you are to the back of the line.
Once you're in the line, we can actually calculate once you're going to get out of the line as well. It's great insight for time management. Time Zone API. This was actually a pretty popular API for Microsoft. We built this for the Windows team who wanted to get off of their own time zones. This gives you, given a point in the world, will tell you what time zone you're in, the offset to GMT, as well as the actual time there, the wall clock time. We just introduced sunset and sunrise times so you can automate IoT scenarios as a part of that too. We have a Geolocation API. It takes an IP address from a computer. That could be a phone, that could be a PC, and will tell you what country that location's coming from.
Mobility is actually an additional partnership we did with a company called Moovit out of Israel. Now you can see there's a partner ecosystem at play. We are very partner-led in Azure Maps when it comes to this rich, deep content and intelligence within the industry. While TomTom brings these rich data and services, the freshest maps, Moovit does the same thing for transit and mobility. In fact, our partnership has extended beyond, and now the three of us are talking. We actually announced earlier this year the first multimodal system that crossed both transit and road graphs. A unique offering coming through Azure Maps. Data storage, if you wanted to store your data in the Azure cloud and use it with Azure Maps, you can use the Azure Maps data storage. There's a myriad of applications.
You can imagine Microsoft has access to quite a developer ecosystem. This developer ecosystem is building a multitude of applications ranging anywhere from mobility fleet and logistics, to IoT scenarios inclusive of indoor maps, facility management, cloud, mobile edge computing. All of this is happening through the power of Azure Maps and location. Spatial analytics and AI. This is also part of what we're seeing as the developers who are building on Azure and using location are building out super rich, cutting-edge type applications and technology. As we go along this journey, TomTom is just coming with us. These are just some of the headlines that have come. I want to always confirm and reinstitute the fact that we decided to not make maps. It was really hard.
I've been at Microsoft 17 years and I was in charge of MapPoint web service. I was in charge of MapPoint, the DVDs. I was actually responsible for killing those. I'm sorry. A lot of people loved those. We had to change our business model. DVDs were going away, and we needed to keep up. I was with the Virtual Earth team when we built 3D and had weekly reviews with Bill Gates. I was part of what became Bing Maps. To this day, the Bing team is still licensing data, ingesting data, and now they're working on how we ingest TomTom data to make Bing Search rich, to make Cortana smarter, right? We're actually using both products, the uncompiled maps for Bing, because they have to ingest it and train rankers inside of the Bing search engine.
We're using the APIs directly for Azure Maps. We're also very enterprise-ready. This is an important facet in 2019. It will get even more critical moving forward. We talk about enterprise-ready. We talk about enterprise scale, right? As our customers grow, we grow. Elastic scale is not a real thing. It's just a concept. With our partnership with TomTom, we've been able to scale. I say we wake up every morning and see who our customers are. Our customers could unload on us 1,000 QPS, 2,000 QPS overnight. We need to be ready for it. That's the enterprise scale we're talking about. The infrastructure that TomTom has built on Azure is actually scalable too. They have to scale with us. We treat them as a back-end engineering team, right? Globally available.
This is accessible to all the customers that Microsoft has around the world. We talk about trusted security services. When I joined Azure Maps and I launched the product, I had to go through a plethora of trials by fire, if you will, 375 different tasks for compliance, security, accessibility, usability, privacy. We are fully supportive and checked down all the boxes for compliance in Azure. The pricing itself, super competitive. I mentioned we can lean on enterprise agreements. As customers are signing up for Azure, they have a menu of options that they can go choose, and Azure Maps is right there waiting for them. They can select it and start using it immediately. They don't have to talk to me. I'm much like Anders. You don't have to talk to me, but you can if you want.
You can just start using the service. It's been a big win inside of Microsoft. Our growth is significant. It's 200% month-over-month in terms of customers for Azure Maps specifically. We only GA'd, general availability, May of last year. The uptake has been significant. We're going to continue to grow, we're going to continue to innovate, and I will continue to push requirements on TomTom to keep their engineering team on their toes as well.
Yes. My first question would be the map making basically partnership that you have with your clients. When was that introduced? How often is it used? Can you give us a bit of a feeling on how big this is within enterprise? Maybe as a measure, out of the roughly EUR 160 million revenues, what portion of revenues do the partners represent in the total pool?
Okay. On the map editing partnership, I would say it is less than a year we introduced it. Of course, you always start with a few to kind of be 100% sure things are working. I think what's important there is that if you want to go that route as one of our partners, you really need to be super serious about it. Meaning, as I said, the training part of it, you have to have certain number of people because otherwise it doesn't really scale for you. That's working fully now. It's implemented, and we have a number of strategic partners
working on it, and we allowed, in this case, to mention that Uber is working and using it, but there's others as well. It's not reflected to our revenues, it's reflected to the modifications we can do on our map database. Basically, making our maps fresher, much faster, and more efficient.
Hi. François-Xavier Bouvignies from UBS again. A quick question on the enterprise, specifically on the indoor data. As we see in the market, Apple probably one of the main rationales for moving into its own maps is to have more information like Street View, probably indoor maps. HERE is offering the same kind of solutions. Is it something that TomTom would look at in terms of is there demands for this in the medium term? Maybe we can ask as well as from Azure Microsoft perspective, is it something that is needed as well from your perspective?
You mean indoor maps specifically?
Indoor map or Street View, kind of additional solution that maybe TomTom is looking at.
Yeah. If you take Street View, Look Around as Apple is calling it. Again, of course, the advantage I explained with our own compound maps with the layers. In this case, Apple have been building Look Around, as they call it, on top of our data, which is kind of similar to Google's Street View. For that specific, you could say, case, we don't have any plans at TomTom at this stage to build a similar product as Apple.
Yep. I can answer it for Microsoft. The Street Side or Street View, we did a collection back in the day, gosh, 2010, [2029], with Virtual Earth and Bing Maps, and we still have the data. It wasn't widely used, and that could be just a reflection of the product itself. We haven't had a lot of customers ask for it. A lot of the scenarios are consumer-based. There are some B2B scenarios or even B2B2C scenarios, but we haven't had many customers ask for it. I could tell you that in 2019, asking for that and the level of investment it would require from our management team is not going to happen at Microsoft. We're very stringent on our finances and keeping track of Azure specifically. Given the requests, their lack of requests I should say, it's not terribly interesting.
The indoor maps is interesting to us. There are some things that if you look across Azure IoT, if you look at the Digital Twins, for example, and smart buildings investments that we've been doing, there's some interest in doing some indoor mapping, and understanding what that technology looks like. We announced Azure Spatial Anchors prior to this. There's some things happening that would force our hand a little bit now. Currently, Azure Maps actually would support the integration of an indoor map. We have proof of concepts and actual customer applications that we've demonstrated at shows. It's natively integrated in, you could take a raster layer and integrate it in our SDK. You could take vector data. You could do these things. Today, we don't have an indoor maps platform. It's certainly something that's really interesting to us.
Yeah. Marc Hesselink, ING again. First question is on, how should I think about the market growth? It feels like the market itself is growing and potentially you're taking a larger share of that market. What kind of potential would that be for you on a growth level? The second one maybe for Chris. When you took the decision to go for TomTom, I assume you also looked at alternatives. What made you finally decide for TomTom as the best solution for Microsoft?
Yeah. As I said, our estimate is that the market is around EUR 1 billion today, but growing. We're crunching through all these numbers as we speak, what kind of growth we're seeing. Of course, it's not always that easy to track these things. I think we'll have to come back on that one until we have some more clarity on it. As I said also that we're relatively new in that space, there's still some things we need to figure out before we can kind of make it more public in terms of the growth. We're also working with Microsoft there to see what kind of trends they're seeing, of course, because we are, of course, helping each other in this partnership. Hopefully in the near future, we'll be able to get some more clarity on that.
Yeah. Our decision for TomTom became very evident. I mentioned that I ran the Bing Maps data ingestion team. Part of my job was to actually look at the data and different data providers. Bing Maps uses a number of different data providers, and we looked at the raw data. We actually looked at the raw data and found it to be better. Regardless of the services that were being used, where most people were testing the actual data itself, and we did ground truthing for this, turned out to be better than what we were currently using at the time. That's one piece. The second was this editing partnership is humongous.
I can give you an example in that it would take anywhere from 12 to 18 months for an edit to be fixed previously. That is from the time of submission to a review process, to a field tru thing, to an actual fix submitted into a quarterly release, and then of course, our six to nine-month release cycle. It was painful. Having the ability to stand up services that will receive our feedback immediately, it gets reviewed by their data team immediately. There's prioritization set to all of this. I guess in the end, it comes down to TomTom's willingness and flexibility to partner. They are just a solid partner for us, and their data is better.
Yeah. Ivan again. Maybe can you elaborate a little bit on kind of the partnership between the two companies? Is it for a very long time or is it that you can cancel it potentially immediately? Also suppose that as you are going to grow tremendously, and potentially kick Google Maps off the throne as being the number one map, will TomTom benefit in equal amount, i.e., if you grow, will TomTom grow along with you?
Yes. What we announced was it's a multi-year agreement we have signed with Microsoft. We definitely have a number of years to partner. That was kind of the whole idea also with, of course, us going to their cloud and putting our products there, and for Microsoft being able to go to it, of course, takes some time, right. That's definitely a multi-year contract we have signed with Microsoft. The other question was, sorry?
If Azure Maps is going to grow the way that probably you guys are expecting it to grow, will TomTom benefit in equal amounts in terms of-
When I get a call from Taco, I need to grow the revenue, I call Chris.
It is true.
That's just a routine, right?
I was wondering where those calls came from. Yes. As we grow, TomTom grows with us. I should be explicit about that because the maps, the search, the routing, and the traffic are all core to Azure Maps. As Azure Maps grows, TomTom's use through Azure Maps grows as well, and so TomTom grows as well.
Maybe from your perspective, Google Maps has been very dominant in the consumer space. What will it take for you guys to actually beat Google Maps?
You want to take that? That is a hard question. Google is a formidable competitor. They've effectively dominated the market for some time, and especially in consumer mapping. They're not impenetrable, right? I think it's going to take a number of years, and a different position entirely, to basically change the way Google has won the market today. If you look and up-level this to what folks are calling the cloud wars, Google's a distant third. We focus on that perspective, and that applies to maps as well. We're part of the first-party Azure ecosystem, and as Azure grows, Azure Maps will grow, TomTom will grow. It's tough. Google's raised the bar to be extremely high. I think it's just going to take a combination of partnerships and a new strategy of how we go about doing maps.
I think the map visualization and what exists today, people are finally starting to appreciate and understand how great it is. There's generations now who didn't understand what it was like to have a phone without a GPS on it, right? They're kind of spoiled in that way, and they actually want more things, which is great. They're pushing the boundaries, and that comes into this invisible part of the map as well. You need to understand what's there without being able to see it. This remote monitoring and being able to control drones and other mobility and ride-sharing scenarios, things like that, where you're not on the ground, you actually don't know, but you need the information to be there readily available.
I think that that perspective and changing it through the IoT lens for us, the Azure cloud lens for us, it actually just kind of permeates into a different strategy that will just be different. I have no plan or intention to go beat Google Maps, especially not by myself. Even in partnership with TomTom, that would be really, really hard to do. I think it'll take a group of us to basically go out and create a new market to change the game.
Good afternoon. Welcome back. My name is Antoine Saucier. I'm running the automotive business at TomTom. I want to give an overview of where we are today, but also how we see the market megatrends as a great opportunity for TomTom in the automotive business. It starts with our product portfolio. You've heard about map software and services already this morning. Yes, we're a map company. Our transactional mapmaking process produces navigation map, high-definition maps, but also ADAS map that are increasingly relevant in the automotive business. We're leading in navigation software with more than 25 years of experience in the user experience and ease of use. Traffic is definitely our leading products.
We're leading on the market in automotive with traffic information. That is basically us being a big data company processing massive amounts of data, and that capacity also works with our products, than traffic, and that is recognized across our industry. How did we start? Basically, we introduced, back in 2009, a product that dramatically changed the landscape of navigation in automotive. Based on the contract with Renault signed in 2007, Geneva 2009, we introduced this Carminat TomTom product below EUR 500, best in class in terms of navigation features, and completely changing the game in terms of in-car navigation take rates. If you look at where we are today with some of our biggest customer, the take rate is between 50% up to 80%, depending on the car lines. By that time, you were talking more 5% than 15%. We completely changed the game by introducing this innovation.
We continued in 2010 by connecting those devices, actually providing this same device, but with a SIM card in it, and introducing real-time traffic information in the automotive industry. That also was a starting point, and we became the market leader that we are today in traffic information. We also introduced first electrical vehicle navigation. That was in the Renault ZOE, one of the very first EV car being produced at mass market size. We launched more recently, ADAS-based Level 2 system in a couple of cars, but also trucks. We also expanded on a geographical coverage. What we announced this year with MG in India, with this Hector car, is actually the first navigation system for us in India, and they also take us to other regions. Altogether, technology coverage and user experience have been the major drivers for our development in the business.
That growth of the automotive business is also supported by the now well-known CASE megatrends. Connected, automated, shared, and electrified. Connectivity is an interesting topic. You could say today in Europe, all the cars are connected. They all have eCall. eCall is a connectivity that is only activated either if you crash or if you press this little red button. If you think about connected services, how can we update our software? How can we update our maps? This type of connectivity is not going to solve the problem. What this connectivity is about is really bringing decent data exchange in the car, allowing for software updates, for real-time maps access, but also to get data back from the sensors being increasingly included in cars.
Since this connectivity is also a cost, trying to monetize what comes out of this connectivity to compensate for some of that connectivity cost. In automated driving, obviously, ADAS and HD content is getting market attention. We also see safety services being now requested. What is the connectivity coverage of this or that particular location? Can we offer connectivity type of map? Can we think about what is called road clearance, which is a service that would tell OEMs, "Well, in this particular location, autonomous driving is allowed," or "You should put this feature on hold because of weather condition, accidents, animals, or any other hazards." We also strongly believe in the value of bridging, over time, the ADAS and AD space and the navigation space. If you look at how our business develops today, we're basically talking to different engineering teams.
Navigation is more in the infotainment engineering department at our customers. ADAS and AD is more managed by chassis control. Ultimately, if you talk about user experience, how do you bring this technology into cars so that people understand it, use it, are delighted with it? It means that you merge the navigation content with the map and you're able to represent your ADAS features or your AD functions at the same location, the same style as you represent navigation. Electrification, a very good case for us as well, because EV cars bring a lot of advantage, but they also bring range anxiety. Range anxiety is a problem that location technology can really help to solve. That is because we know about the map. I'll come back to that. We can bring electrical vehicle services.
We can deal with charging stations, and over time, organize that not only you know where the charging station is, but when you reach that charging station, it is available because you've booked it in advance, and the transaction can also be taken into account by the car. Those trends can dramatically boost the navigation take rates in automotive. If you look at what has happened in the previous years, we've been growing steadily but slowly from 29%, 30% and plus. If we would project that trend towards 2030, we would be around 45%. We strongly believe that whether it's the connectivity, moving towards more automated driving cars, shared mobility or electrification, all those trends generate real use of in-car navigation, real demand for something, some system that addresses those features in a proper way, and that can really boost the take rate in navigation.
This will only happen if the expected user experience from the drivers, from the end users, as Harold mentioned already this morning, is taken into account and is actually addressed. I think we're at a turning point today where most of the in-car systems have limitations because the software is not decoupled from the hardware. Therefore, it's outdated when it's launched, not easy to update. Map is not online and also not easy to update. Ultimately, the resulting customer experience is not at the expected level. That has generated reactions from some of the OEMs. You've seen bigger screen coming up, different layouts, vertical screen, horizontal screen. I could say Byton is going horizontal and vertical in all directions, or Daimler trying to better integrate the cluster display together with the central stack.
Ultimately, what is beyond that, so how can we help those OEMs, is more behind the screen and is in helping them in bringing the connectivity, connecting those devices, developing the software more independently from the hardware and moving into navigation as a service, much more than navigation as a product that would be developed exactly at the same pace as the car is being developed and not updated after its delivery. The fundamental change is really in this software decoupling from the rest of the car and update through connectivity. How can we deliver that ourselves to OEMs? How can we support this transition? Well, first of all, it's in navigation. What we call best-in-class navigation is not only what I already mentioned, so fresh software from the start and updatability, but it's also getting data on the actual usage.
That was also something that we've demonstrated in Frankfurt Motor Show together with Microsoft, getting data from the actual usage. What are the features that ultimately drivers are using? How many clicks are they pressing on the screen? How far do they go? When do they drop this or that feature? What is the pattern that they have in their behavior? Learning all of that can dramatically change the view we have on how navigation works and how do we want to bring this or that feature back in the car. Autonomous driving, so ADAS or AD, also needs to be better integrated into that experience. We're ideally positioned to deliver that. EV, I'll come back on that later on, I think that's also a very good case. Also, connected to the digital life. I was talking to Chris last night.
He was telling me how he implemented Alexa in his car. I was very interested. I thought this would be a fantastic software experience. It ended up ordering a small Alexa box on the web, plugging that on the cigarette lighter or the USB plug, bridging it to your smartphone, Chris, correct, to get connectivity, bridging this device to the car through Bluetooth to get the sounds over the speakers. I thought there must be value in having Alexa in your car. If you want to go through all of that process, that is what Chris explained as well, that once all of that is done, he has access to everything in his house and his life that is Alexa connected. There's a real value in bringing Alexa in the user experience of the overall infotainment.
It changes the picture in terms of voice recognition and so on and so forth. definitely the value is not there in terms of how do you install that in your car at the moment. that tells us there's market demand that needs to be taken into account and addressed in the way we develop our product. overall, whether it's voice recognition, destination entry, digital life integration, ADAS or AD integration, it's all about user experience, and it only works if you move from the current onboard to more online, more connected. by the way, you don't want to address Alexa only because then if you're a Google Home customer, you're not satisfied either. it means that we're able to have our software evolving, and taking into account all those different offers from end users.
If we zoom into some of those opportunity, I'll start here with the connected services. We've been very successful in that space. I mentioned already traffic information. You see a fantastic growth forecast here. I think navigation system without proper traffic information in the future will not survive, and therefore you see a strong growth for traffic. You also see that the number of services that will be integrated in the packages that we deliver is increasing. Right? Where we started with traffic, we quickly also introduced speed cam, parking. I think today we're looking for parking spots, but tomorrow we're going to look for parking plugs, rather. That continues, ADAS and AD also requires more hazard warnings. Weather is also important in certain conditions.
The moment you go into EV, but also parking, it's the same space, we're interested in, how do I make sure from user experience point of view that my charging station is available or my parking spot is available? Therefore, we talk about the capacity to book those things and then to pay for them, and just making everything smooth along the experience for the drivers. The level of services is increasing, but also the number of services that are in these packages. On AD, we've been also quite successful. I think you've heard a couple of things on the strength of our maps. I think that also has played a role in the businesses that we've won on the HD map market. It's a new space. There is no legacy, there's no existing market share.
We see our customers deep diving in the technology and then making their choice and a significant portion of them have decided to go for TomTom. I think that's the best proof of the quality of our technology. It's true for HD, it's also true for ADAS as an extension. ADAS attributes have been in our maps for a while, but they're increasingly being used by customers in ADAS context, and that can be driven by new features coming in. Also, Euro NCAP in Europe is driving for speed limits and speed assist system is driving an increasing adoption of ADAS data. Also Europe is talking about regulation around those intelligent speed assist system. That is a very good opportunity for us. We also expand our portfolio of services. I mentioned that already. That's the road check or road clearance type of product.
Again, we strongly believe in ultimately the merge of navigation and ADAS/AD features into one customer experience. Moving towards EV. This is where our complete portfolio completely fits the market. It starts with maps. Of course, you drive your EV on roads. You want to know where the charging stations are. We will also use the consumption model coming from the car, and therefore, we need curvature and slopes or gradients from the map. There's a lot of map data that has specific use in the EV space. Charging station. I think we have still 23 different types of plugs worldwide, so making sure that you have the right charging station is an important point. Availability of that particular plug is also crucial if at the end you reach the station with a low level of battery.
Again, you want to be able to book, you want to make sure the station is going to be available and you're going to charge safely. On the software side, we're looking at what is called multi-stop routing. I think in this example, it's a drive from Amsterdam maybe to Nice or Marseille. Quite a drive. There's a bit of strategy to decide where do you want to stop, how long are you going to stay, and that depends on what is your battery type and what is the best consumption charging cycles that you want to follow. There's a lot of interaction between the OEM on the consumption model side, but also the configuration of your journey. What is your destination? Also, temperature will have to be taken into account if you need climate control or heating, or not.
Also, at some point, your digital life is going to come in. If you prefer to stop at some friend's location instead of spending an hour or so in a service station. All of that comes together again in the resulting driving experience in those electric vehicle cars, removing the range anxiety from customers. Finally, this EV experience is not only limited to the in-car. Most of your trips, you prepare them. You want to check before you depart. You want to make sure your car has enough battery to reach your destination. All of that should happen on your smartphone or on your laptop screen. Nobody understands anymore that there is a clear separation between what happens in your car and what happens in the rest of your digital life.
As you can see, EV is a fantastic use case for the complete integration of our product portfolio, but also to push OEMs to better consider the integration between smartphone and in-car experience, and that's also where TomTom can support. Ultimately, it should look like this. Okay. Last but not least is the digital life part. I already mentioned the example of Chris. I think that is something where also we need to help OEMs to open up and make sure whatever is your digital life content, if I'm a Deezer user, I want to get in my car, be able to use that without having to do all sorts of clicks and plugs and things that are mandatory today. Voice, I think, is also a fantastic example and use case for us to get OEMs to change.
They've been trying to get voice recognition working in cars. It has never really been amazing as an embedded feature. It completely changes when you move to online. The result of that is the Alexas and others are obviously looking into what's happening in the automotive. Again, you cannot be exclusive to one or the other. You want to be able to address that, whatever your user is using. I think that's part of what we need to support as well. In terms of business updates, in the past month, we've been announcing quite a bit of news. Navigation, we launched with Jeep, Renault, and Nissan brand. I was talking about 2009, that was the first Clio for us to go with our navigation. 2019, 10 years later, the new Clio is also introduced, also in Geneva, and still with TomTom data in it.
I think that's quite impressive. Our customers, they tend to stick to our technology, continue to trust our teams, and our capability to move them to the future with their customers. We also introduced IQ Maps in India. I already mentioned that IQ Maps basically make sure you drive the freshest map available wherever you go. If you're in India, it's quite a big amount of data, you don't want to have the full of India updated every time, but just the portion of that map that are of interest to you, that is what IQ Maps is delivering. We're also preparing a big launch for a big OEM in North America. I think what you've seen is over the years, we've completely changed our game in North America. We're delivering Nissan, we're delivering Subaru, we're launching with another OEM.
Our business in the U.S. has dramatically changed. Connected services, Volkswagen renewed our agreement together with them. MG, I already mentioned. We won a couple of major deals. Traffic has been a fantastic door opener for us in the automotive industry and continues to drive discussions, but we also deliver those services as part of the full package. EV, as I mentioned, we pioneered that technology with ZOE and Renault. We've extended that to Nissan LEAF. I think with that, in terms of volume segment, we're definitely leaders. BMW also trusts us to deliver online this type of service to their cars, also routing. For Jeep, they introduced hybrid versions of Renegade and Compass, and those are powered by TomTom technology. In the automated driving space, it's been mentioned already, we've won multiple OEMs.
Again, I think technology and trust have been the key driver for those decisions at OEM. We've launched AutoStream. I think Willem will tell you everything about what that technology is about, but basically how do we distribute maps, HD maps into cars. With our ADAS map, we're already in business and our customer base is growing fast. We have now more than 1 million cars and trucks driving around on an ADAS map from TomTom. I think that's it for me. What you can see is that through our portfolio, our capability to bridge together navigation, which has been our world for a very long time now, and ADAS and HD features that are now fastly developing, we're best positioned as company to help customers going through this transition from onboard to online, securing that they can deliver the user experience that customers expect. Thank you.
I'm Kees van Dok. I am Chief Product Officer at TomTom. I've been with the company for about eight years, and I've always been deeply involved in the end user experience of both automotive as well as consumer products. I would like to spend the next 15 to 20 minutes to give you a heads up, a view of how we feel, how we think that customer user experience will evolve, and also how we think we can take a differentiating position in navigation in comparison to the competitive field. Before I do that, let me zoom in a little bit closer on what Antoine already described, that the automotive industry, and especially software in automotive, is in times of disruption. There are a couple of key kind of enablers that are happening in that space.
First of all, I think we see automotive software makers struggle, if not failing, to meet customer expectations. Customer expectations have rapidly evolved because of mobile phone usage. This is a quote from J.D. Power. J.D. Power is a well-respected automotive consultancy, voice of customer company in the U.S. They reported last year that about 20% of new car owners with navigation in their car actually never uses navigation. An even higher percentage is stopping the use of embedded navigation in the first 90 days of their new car usage. Instead, people flock to their phones. They trust phones for navigation and other kind of digital life experiences that they expect to be able to continue to use in the car when they move from all these other contexts into their car.
That obviously is because the experience delivered on phone is by far superior to what you get in most cars in terms of up-to-date software, in terms of fresh content, speed of user experience. That's the first point. Another point is that the paradigm shift that happened in the mobile phone industry about a decade ago is now happening also in the automotive industry. Where you see the replacement of a lot of feature value and end customer value through hardware controls, shifting to a software first way of delivering that value is a paradigm shift that is now happening in automotive. That delay is obviously because of the nature of the industry, where development cycles of cars are taking a very long time, but also because these disruptive elements are happening.
What you see in the car space is that the environment becomes more familiar to what users expect from their daily life experiences, right? The car becomes more like how you interact with your phone, your TV, with your home automation software. We see a vast reduction in hardware controls and buttons and levers and controls, as you see on the left, and a much more software-centric user experience driven by voice, by touch, and by a limited set of steering wheel controls. Obviously, a third disrupter is the entrance of Google in the car. Google Automotive Services went from mobile first to Android Auto to becoming really embedded in the car, bringing the, it was quoted before, phenomenal Google Maps experience in the automotive environment, which is obviously a big shift from where most car makers are today.
The question is, how are we as TomTom, how are we going to deal with that, and how are we think there is a path forward in our navigation software user experience? We identified four different pillars which we base our strategy on. First of all is to be really mobile competitive and to embrace the best practices of the phone industry and the mobile industry more applied to our own full stack of navigation products. Rather than having embedded maps and having software that runs in the car, we're currently in transition of shifting all of that to an online reality and make sure that part of our navigation software is all running in the cloud. We had a lot of proof points this morning about maps moving to the cloud.
Embracing that paradigm and being online first is currently happening, and I will show you some proof points of that. Integrating with users' digital life is part of that. How can we bridge the value and the services that people trust on their phone, how we can bring that into car and make that transition as smooth as possible. Connectivity is obviously required, and that has always been a difficult discussion with automotive makers. Broadband connectivity delivers the best user experience, but it also comes at a cost. Is that cost being transitioned to the user? Is that taken by the OEM? It is a difficult position to move forward, but it is essential to deliver a strong user experience. Lastly, being voice centric is key. In the car, safety is what we are all about.
Interacting with software and features through a lot of manual actions is quite unsafe. We're building, we're creating, we're preparing ourselves for a voice first user experience of our navigation software. The second part of the strategy is to really differentiate in navigation itself. Rather than relying on mobile-derived, smartphone-based navigation only, we feel there is a position for navigation to be more integrated in the car itself. It was mentioned a few times, like taking ADAS as an example, and I'll show a few more examples of that later in the presentation. Like taking all of the data that the car generates, the sensor data, what the car is observing in its environment, are perfect data points and feedback mechanisms for drivers to integrate in the navigation experience. I'll show a few examples of that.
Thirdly, we feel we need to also zoom out from navigation only. Currently, our navigation software is often part of an in-vehicle infotainment system where many other applications are kind of collected in order to deliver an end-to-end user experience. All those applications often have their own user interface paradigm, their own way of working, their own way that their information is being organized and being accessed. We feel strongly about creating a more holistic way of interfacing with that software in the car. How you control software, how you control it by voice and by touch and by other ways of input. We want to create a view where moving from, say, communication to entertainment to navigation to controlling features in the car feels much more the same across all those applications.
That's obviously zooming out from navigation only, but we feel by only focusing on such a more holistic user experience, we can actually deliver a more delightful and better user experience for car drivers. Lastly, we don't do this alone. It was mentioned, MCVP is an important technology partner. Having a robust, scalable cloud platform that enables a great in-car user experience where you have authentication, good data collection in the cloud, lots of services with high quality and high speed delivered into the vehicle is essential. Also on the voice assistant angle, is not technology that we would ever develop within our own company, but relying on the right complementary partners to provide that more holistic user experience is the way to go. We're developing those partnerships as we speak. Close collaboration with OEMs is another important angle.
We are, I think, OEM friendly. We work with OEMs, we have deep engagements and partnerships with OEMs to together build solutions. For our next wave in navigation or IVI, that is not any different. Understanding OEMs needs and requirements when it comes to customization to create a user experience that feels authentic to their brand rather than a brought in solution by a third party is an essential element there. Now zooming into a couple of proof points. We're moving our navigation stack to an online first reality. TomTom AmiGO is the first proof point of that. We're developing, as we speak, a lightweight navigation application where all of the navigation software runs in the cloud, runs online.
All your maps, search, and routing, and all those aspects around navigation are coming from the network, and there's only a small, lightweight application footprint happening on the client. What we're trying to do with AmiGO is to extend the already great presence that we have with end customers. We believe end customer and the relationship with end customer is the way to ultimately deliver the most delightful experiences possible that we then bring to our automotive customers. AmiGO is in open beta on the Android platform. It was released a few weeks ago. If you have an Android phone and you have the TomTom Speed Cam app on your phone, you can upgrade to AmiGO to experience this lightweight turn-by-turn yourself. The second initiative that we started is product concept development around what I mentioned, this broader scope of IVI.
We're currently doing a product development, concept development, around an end-to-end infotainment or in-car vehicle infotainment system. Again, also taking a view on how communication would work in the car through your phone or through conference calling. How entertainment would work. The ability to play music, either streaming from the internet, streaming music from your phone, or from connected music services. How that all kind of interplays also with navigation without any application switching and context switching. We're developing that view in order to, again, have a view towards OEMs of what we feel an integrated experience could look and feel like, but also to have a product that we can test with end customers and test with OEMs to see what kind of features would really stick and work for people. Let me dive in a little bit into the navigation opportunity itself.
As I mentioned, we also feel that in navigation software, based on the stack of technology that we have from map to ADAS, to autonomous driving HD data, to live services, we feel that on top of that stack, we can actually also play a larger role in the in-car vehicle experience by bringing those technologies together into a driver experience that is more front and center than what navigation is today. I'll show you a few examples. This is a demo that we built a few years ago, a real-world working proof of concept. It's based on HD data. This is an HD map rendered in 3D. We take the camera position from the car into account in the navigation experience.
We see the accurate lane position where the car is in, and we give kind of this yellow bow wave nudge to move the driver into the right direction. Rather than giving navigational instructions like, "Take the right at the next junction," we can actually much more smoothly guide the user through the right lanes in order to make their navigation experience much more natural and much more fluid, compared to today. That can only be done by taking not only the map data from our map service and integrate it in the navigation experience, but also take information from the car sensor data locally in the car and bring that together into one unified user experience.
1.2 miles.
A proof point that we're also shipping this kind of technology already. This is not only concept work, is in our most recent version of the TomTom GO Navigation application, which is more like our embedded maps premium navigation application, where we brought this whole concept of moving lanes into the user experience. This is actually one of the differentiating features in that application. Again, by bringing in sensor data, we can enrich that even further. That's also trying to address a real-world problem, because if you take a driver in a modern car today, there are quite a lot of different places in the car where they get their feedback from in order to take navigation and driving decisions. For example, they may have a screen or a phone with navigation software running on them.
They may have a next instruction panel right in front of them or in a center stack display. There may be all sorts of warning mechanisms, like gentle warnings coming from a live service. There can be specific hardware-based features like collision detection, blind spot detection that can sit in a mirror or in the display right in front of the driver. There can be collision warning, little LED screens in yet another location or on the steering wheel. We're getting all sorts of cryptic visualizations about more advanced features. For example, this is the user feedback for L2 driving, where the car kind of sticks to the lane and keeps the distance to the car in front of it. Sometimes it's even more cryptic, little icons that light up somewhere.
There's the outside kind of world information that the user also needs to kind of digest and understand in order to take decisions. That can be a simple message that can also be part of the ADAS map, or it can be a much more complex message that is quite difficult to parse when you're driving by quite fast. We believe we have a position to take all that information and kind of bring it all together into a unified driver-centric user experience, because this whole hodgepodge of features that is sprinkled across the car feels a bit like a bunch of ingredients where the chef and the recipe is missing. We think we can play a role there.
This image brings a couple of those items together, where we see a view on the road ahead, and that is not an unfamiliar view from what users expect from navigation today. It has a lot more information. This is not only a green line that gives you guidance, but it's also an HD map that gives you a real-world view about what you see out of the windshield. It gives you sensor feedback. There is a car on the right of me that is trying to merge to the left. The car actually observes that maneuver and warns me for it. It's a much richer experience. We bring in this kind of setup together, our map, obviously, address data, routing and guidance as we know it. It also brings in information that is coming from the traffic service, live services.
Doing a gentle warning with a jam that is around the corner is integrated. Also, as I mentioned, this sensor feedback from the car itself. Imagine my car detecting another car in a blind spot, or there's a collision warning, or as I shown before, this lane position thinking. We bring all of that into that same visualization. That goes all the way up to the deeper levels of automation. This view can also guide users towards autonomous driving in the end. We know that the adoption of autonomous driving is very much a user trust and a safety issue, like users will need to get to trust their car to make the right decisions. That all starts with building trust through visualization.
If the car can explain very clearly to the driver what's happening outside, what kind of decisions it's going to make, how it's going to move, what kind of maneuvers it's going to make, the better that trust building will happen. This is not just about next-generation navigation, but it's also a view on how that can expand to an autonomous reality further down the road. These are a couple of visualizations of how we think we can bring that to life. Imagine this is the cluster display in front of the driver. Here's an example of a collision detection integrated into navigation. Here, another example of a blind spot detection.
I'm instructed to move to the left, but there's actually another car to the left of me, so it's not safe to take that maneuver, and as soon as that car has passed, I get the green light or the blue light in this case, and I can move over to the center lane. Another example is, as I mentioned, level 2 automation. The car sticks to the lane it's in. We visualize that by a couple lines next to the line that I'm currently in. As soon as it needs to change speed because there's another car in front of him, again, we use that same kind of rendering environment to visualize that context. It's a little hard to see, but there's a white car in front of him and distance is being checked. Again, also for online information.
If we know there is traffic around the corner and you're approaching it too fast, we can do our gentle kind of warnings of overspeeding in that exact same context. Lastly, an example around how that could even traverse into autonomous driving. There's an autonomous driving zone coming up. The driver is informed about that zone coming up. There's the handover to the autonomous car system. Again, all that information is being brought into one place. Rather than sprinkling dialogues and messages and beeps and notifications all over the place, we believe that that view on the road of the next couple of hundred meters in front of you is going to be an area that we are very familiar with.
We've been rendering that view since our very first PNDs, and we believe there is a strong possibility to naturally grow that into a next wave of navigation. I still remember very vividly the first time I stepped into an autonomous vehicle. The engineer opened the door and said, "You go sit behind the steering wheel." I sat there, and he instructed me to drive to a highway stretch, and he said, "Now you press these two buttons, and then you can let go of the steering wheel." That's when the magic happened. That's when I knew this was going to be something. This was fantastic. Within minutes, I was talking very excited to the engineer in the back, who urged me to look to the front of the car again because it wasn't that safe yet that I could be talking to him all the time.
Autonomous driving, if you've experienced it, you know it. It's going to be huge. The benefits of it are just enormous. I'll talk a bit about autonomous driving today. I'll ask one of our partners, Erik Koeling from Zenuity, to come on stage to explain some of the elements that we don't make. I dive a bit deeper in what the role maps is, how we make those maps, and I guess you're dying to hear what the market size is of this opportunity. I'll dive into that as well. A few months, actually, after that experience in that vehicle, I was leading the autonomous driving unit of TomTom. Autonomous driving, it is happening. Yeah. The societal benefits are so big. The investments are so big, so huge. It is happening. The safety advantages are enormous.
There's every day, 3,000 people dying across the world, and a multitude of that are getting injured because of vehicle accidents.
It gives us comfort, right? If we have truly autonomous vehicles, we can actually look at our smartphone, we can work in a car, we can sleep. It gives us back time that we're losing out on now in our commute or on holiday. Wouldn't it be great if you step into your vehicle, you drive it to the highway, you press activate function, you drive to the south of France, and you turn around your chair and play games with your kids? That's the comfort part. The efficiency part, it'll change our cities. It will get to better fuel, lower fuel consumption, all of that stuff. It's happening. What do you need for it? Well, you need four things to make it happen. First one is mapping. No surprise there.
It's a very important element, and the higher the automation level, the more important maps become. You also need sensing, so you need sensors like cameras, you need laser scanners or lidars, you need radar to see what is around you and compare that to the map. You need a driving policy, meaning software that takes the decisions to go left or right or to overtake. You need actuators that actually replace your foot and your hands on the gas pedal and the brake and the steering wheel. I'll go a little bit deeper into mapping, but I'll first let Erik Koeling from Zenuity talk about all the other stuff that it takes to make this happen. Please, an applause for Erik Koeling.
Thank you, Willem. A very good afternoon. I'm Erik. I work at Zenuity. As Willem said, self-driving is a true Can I take this one? Yes, okay. I can confirm that autonomous driving, self-driving car is a truly transformative technology because it will change the way you operate a car, the way you own or maybe not own a car. It will affect the road transportation system as we know it. That's a good thing because, as Willem said, the road transportation system as we know it is not efficient and is very dangerous. A lot of people get killed in road traffic, and we believe that's unacceptable. Because this is such a transformative technology, our company has been founded a little bit more than two years ago.
We were created by the safety leaders of the automotive industry, that is Volvo Cars and Autoliv as a big tier 1. Two years ago, these two companies realized that self-driving and advanced ADAS systems is a crucial part to remain safety leaders in the future. Developing this technology not only requires different working methods, but also completely new business models. The automotive industry is very much organized around bending metal and all the timelines that it has for bending metal. This technology is much more software intensive, much more data intensive, and that's why we decided to carve out the development of active safety software and self-driving car software from Volvo Cars and from Autoliv and build a new joint venture around it. A pure software developer focusing on this type of technology.
I can say that Autoliv has since then spin off their electronics business into Veoneer, and today Veoneer is 50% owner of Zenuity, and Volvo Cars is 50% owner of Zenuity. What we do is software development for sensing, driving policy, vehicle control. We partner with TomTom when it comes to mapping. We already had the workshops before Zenuity even formally was founded. We were sitting together to understand how can we build together, build a complete stack for vehicle automation. We're working on that full speed ahead since two years ago. The way we do that is, we have made choices in how we want to develop this technology, because when you talk about a self-driving car, it can mean a lot of different things.
Some people may think of self-driving trucks or robot taxis where there's not a steering wheel at all, while we focus on automation for, let's say, privately used vehicles. That is vehicles that still have a steering wheel. They just sometimes drive manually, but sometimes drive autonomously. We believe that that is a very attractive way to enter the era of self-driving cars because ADAS is happening in almost all vehicles. NCAP is requiring that for all vehicles to get a good rating, and we also see from the OEMs that a new vehicle just has to have at least a camera and ideally more than that.
By being able to build an ADAS business in parallel to a self-driving car business, we can move step by step forward and deal with the uncertainty around creating revenues around self-driving cars, as opposed to doing robot taxi vehicles where you don't create any revenue until you completely solve the automation problem, which is very challenging. We instead do this hand in hand with ADAS such that we can build a business over time. Before I joined Zenuity, I had been working at Volvo Cars 18 years in the field of active safety, pioneering in the early days, adaptive cruise controls, automatic emergency braking systems, pedestrian detection systems, and things like that.
During those days, we actually were never allowed to use a map, because an ADAS system is being offered independently from the navigation system to the end customer, and we could not get the guarantee that there would always be a map in the vehicle. Fortunately, that time has passed. We see that you cannot build a premium ADAS system without having a map in the vehicle. We also see that if you really want to have a leading ADAS system the coming years, you need to have an HD map in the vehicle as well. On the screen, I try to depict what we believe we need to develop the coming years to move forward.
We will, at the end of this year, early next year, launch the next generation of active safety systems at the Polestar 2 vehicle, new collision avoidance features, Level 2 vehicle automation systems. During next year and the year afterwards, we will do hands-off Level 2 automation, where we also use a HD map. Moving forward in 2022, we will start with unsupervised automation. That will, for us, be the first products where we can say to driver, "Now you can do something else behind the steering wheel." That's a very big task. It's very challenging, that is what we will do. Moving forward, we believe that automation, for us, it will start somewhere in a limited so-called operational design domain. For us, it will be traffic jam driving towards highway driving.
Over time, we will address other use cases that are important for the end user of a privately owned vehicle. Typically, parking is one of those scenarios that we will target. We will do ADAS and AD hand in hand, and we will move forward like that. Building these types of systems, as I said, is very challenging, and probably challenge number one is safety. This technology has the potential to get us to Vision Zero, no road fatalities at all. We also realize that we can introduce new risks into the system by introducing these type of technologies. That is, automation can go wrong, computers can crash, sensors can get blocks, map may be wrong. All these kind of aspects we have to deal with moving forward. That is what we do.
Building these systems, in order to do that, consists of a few layers, and I can use the layers that Willem mentioned. The first one of that is sensing. We have to understand what is happening around the car continuously. An important part of that is the cameras, the radars, and the lidars that we are using. Here you see some results of the 360-degree camera system that we do, where in each detection, in each frame, we detect, in this case, other dynamic road users. In a similar way, we do detections of lane markers, free space, barriers, signs, and a lot of other stuff. This gives us frame by frame a lot of information of what's happening around the car. We can build in a little bit of short-term memory as well by tracking these objects over time in between frames.
This doesn't give us long-term memory. The car doesn't know that it has been there before. The only way to get long-term memory into the car is to actually use maps. Sensing and getting the perception environment is really important, but we also want to know where are we in a map. We have to know that for a couple of different reasons. There's the obvious ones like when you do an automated vehicle, it has to take the correct route, right? You have to be in the right lane. That kind of planning we take from a map. We also do some operational planning based on a map. When we know things like curvature, we can adjust the speed, and we can adjust the steering so that you can do smooth driving.
Also other things which are actually very safety critical, like localizing yourself in a map to only activate the automated driving system when you are at the right location is very important. As I said, one of the first applications that we work on is highway pilots, highway automation. Thereby, the driver can do something else behind the steering wheel. We have to be really sure that we actually are on a highway when we activate that and not on a road that is parallel to that highway. Determining that you are within your operational design domain is a very critical thing, and that is an information where we use the sensors of the car in combination with the map.
We're constantly comparing what the map data says with what our sensor data says, and only if we have very high confidence that they say the same thing, we allow the system to be activated. If we're uncertain, we decide to not activate it. There you can see that also in those kind of safety critical events, the map plays a very important role. As I said, even before the start of Zenuity, we started to work with TomTom at an R&D level to do, let's say, mutual learning. Doing this kind of vehicle localization, knowing exactly where you are in a map, in which lane you are, what's in front of you, what kind of curvatures, is there a barrier to the left, is there a barrier to the right, that's a new technology, and we're still in relatively early days.
Together we've been exploring these kinds of models. What is it that a sensor actually can see? How can you represent this in a map? How can you maintain the freshness of the map over time? Those are things that we've been working on together since the last two years, and we're making really good progress in that. Also crowdsourcing is a very important area. Crowdsourcing is the technique where you use vehicle sensors to detect, for example, landmarks or lane marks or signs or something else, where the car detect them, you send them to the cloud, you compile that into the map, and then you can send an almost real-time map back to the vehicle. This is a technique that I believe is going to be super important in the coming years.
We have built early prototypes together with TomTom, where we use Zenuity vehicles with a Zenuity camera that detects traffic signs. We detect that, we kind of compress the data, send a small package to the cloud, and from the Zenuity cloud, we send a Roadagram to TomTom, who then integrates this into the map and can update the map with a traffic sign that we detected, but it was never seen before. This kind of mechanisms will be built into the fleet of cars the coming years, allowing you to have real-time maps available for highway automation and for ADAS systems. For ADAS, this is important because it can really improve the performance of the ADAS systems. For vehicle automation, this is important because it will improve the availability of vehicle automation. As I said, we can only automate when map and perceived reality correspond.
These two are key areas that we have been exploring, that we will explore, and we will learn together moving forward. Those are key techniques both to make driving safer and putting ADAS technology in more and more vehicles. It's also a technology that is going to be crucial to enter the area of self-driving vehicles. Thank you.
Thank you very much, Erik. That really helps in understanding the wider picture. If we now dive a little bit deeper on maps. Which maps do we make for autonomous driving? Well, they've been mentioned today. We make the ADAS map and the HD map, and I've plotted them here against the six levels of automation, right? Level 0 through to level 5, where level 0 is that a human controls everything in the car, but it can get warnings from sensors and maps, like you're driving over the speed limit, to level 5, where you can tell by the symbol, where you don't need a steering wheel, where you don't need a gas or brake pedal. It's all gone. That's still a little while away. We're now in mass production at level 2. Tesla Autopilot is an example, as is the Volvo system.
They can all keep lane and adjust speed. They do, in technical terms, lateral and longitudinal control. That's a level 2. A level 1 is typically an adaptive cruise control, where you still have to hold the steering wheel. Our ADAS map is used for the lower levels of automation, level 0, 1, 2. Our HD map is used for level 2, 3, 4, 5. You see that there's an overlap in level 2 because we actually see both. We've sold both. We have level 2 systems on the road with an ADAS map. We have level 2 systems coming up with an HD map in production vehicles. To explain the difference very briefly, an ADAS map, it has been mentioned many times in the presentation. An ADAS map is a road level map.
If you have a five-lane highway, it's seen as one road. It'll have a lane count. It'll have one speed limit. It'll have one curvature, one gradient or slope. Whereas an HD map is a centimeter level accurate lane level map. Actually, you know everything for each lane. If there is an exit lane, then that exit lane can have a separate speed limit, like in Germany that happens. It's a road level map versus a centimeter level lane level map. What's the role of maps? There are three roles that maps play in a car. One is perception, which is shown here, and I'll explain the picture. One is localization, and one is path planning. How does a map help a car understand the environment around it? Well, here's one example.
Let's say this is a picture from a front-facing camera in a car, the car now has to decide, am I allowed to drive or not? It sees this bunch of red lights, it has to make up its mind, right? How do you interpret this picture? Well, a map will help the car determine this. It'll tell you in which lane are you today, are you now. If you're in that lane, it's that traffic light that applies. It can tell the camera, actually, look at that set of pixels, because that's where you're going to find the traffic light that you need to read, that applies to your particular position. It'll tell you, actually, it is okay to drive, right? It finds the green traffic lights. Localization is the second element, the second function of an HD map in a vehicle.
Localization is needed because GPS, which is traditionally used to put yourself in a map, is not precise enough. It's not good enough. You need more. We add special localization attributes to the map, things that a camera can see, or that a lidar can see, or that a radar can see. We add that to the map, and that can be compared to sensor readings. If you then do the math, you can find out in which lane you are. Which is kind of important. Once you know in which lane you are, you can also determine in which lanes the other vehicles are. That is not always trivial to determine, but quite important for decisions, right?
If you're driving much faster than the vehicle next to you, like what would happen in Germany, and you're entering a curve, then it's quite hard to see whether the car ahead of you is in your lane or not. With a map, you can get a much greater certainty that you can continue speeding, and you can safely pass the vehicle that is driving slower. That's an example. The third element where an HD map comes in is path planning. Erik already mentioned some examples. A simple example is if you're on a five-lane highway or a four-lane highway here, and it's a lot of traffic, then you want to know kilometers ahead, actually, that you need to start merging into the lanes to the right. That's where an HD map can help you do that type of path planning, right?
It plots all the vehicles, or the system can plot all the vehicles around you on that map, and it can see very far ahead, much further than the cameras can, because it knows in two kilometers you'd need to be on the exit lane. That was how a map was used in a vehicle. I've spoken about autonomous driving and the pillars of autonomous driving, and then the role of HD maps. How do we make those maps? Just very briefly. We master this because, as you have heard Alain say, we can build on a very long experience and a very wide experience as well, right? We master all the technologies that are needed. What does it take? It takes centimeter-level accuracy. It takes, ultimately, millions of sources.
Sources can be our own survey vehicles or MoMa vehicles, mobile mapping vehicles, as Alain was calling them. They can be Zenuity cameras, can be other cameras. In the end, that will add up to millions of sources that are all differently calibrated and are made by different vendors and create different types of data. Millions of sources, and we have to do that in real time, as fast as possible, because the faster we can reflect changes in reality back in the car, in the map, the safer the whole experience will be and the better the user experience will be in the end. If we go into an example for each of these. First this. This is not a map. This is somewhere halfway through our production process, right?
We get a bunch of raw sensor data in, then somewhere halfway we're here, and then we abstract this into a map. Now, why is this special? If you look at this, you can see that you cannot get this from one drive. You cannot send out a survey vehicle or any other car and drive this in one go and have it all sort of fit. It takes many drives to do this. You can also see that there are no shadows in this. If you have many drives, you need to combine them, right? That needs to be at centimeter-level accuracy. You can't have two lamp poles right next to each other where in reality there's only one. You can't have two traffic signs and all of that. That's what we're able to do at scale.
This is really a leading edge, state-of-the-art application of technologies. This has been done scientifically on small areas, but never at the scale of the world. We're doing it at the scale of the world. Centimeter-level accuracy. There's millions of sources. Here's one example. In the top left, it's left for you guys as well. Top left, you see a video image from a car driving. In this particular case, it is in Japan, in the area where the Olympics will be held, Odaiba. You see a car driving there with a camera, and it is registering information, much like Erik was also describing. It is registering road edges in purple. It's registering solid lines and dashed lines that are on the drivable surface in pink. Below that video, you see how that is stored all the time, and that creates a Roadagram.
Yeah. That information, if you combine it all together, essentially creates a Roadagram, and that's what we do with multiple partners, including Zenuity, to close that loop and get information about what's changing as quickly as possible. Millions of sources. Finally, there's the real-time element of it. One part of being real-time is that you actually need to be able to stream maps to the vehicle, and that's what we do with AutoStream. It's like a Netflix or a Spotify, but then safe for maps for autonomous vehicles. Here's what it does. We're planning a route. This all happens on the navigation map. We're planning a route from Las Vegas to San Francisco. That route is planned and is handed over to the autonomous driving system. This sort of takes you through the configuration. Yeah.
A driver in a car would not do this, but this is how you would configure what data is then streamed, because bandwidth is a concern. Obviously, we want the road elements. We want traffic signs for localization. We call our localization products RoadDNA. The traffic signs we're going to put in, RoadDNA signs, so that they can be compared with the camera and we can localize. We are going to put speed restrictions in, because we want to know that we're not autonomously driving over the speed limit. We're going to put another localization attribute in that works particularly well with radar, which we call RoadDNA Roadside, which is that sort of Minecraft blocks next to the side of the road. We're going to put in jam tail warnings, live data, plus explicit curvature and gradient.
An HD map is so precise that you can calculate any curvature you want, but sometimes OEMs want that to be explicit, so that they don't have to calculate it in the vehicle. We're going to start and download only the relevant tiles that we need. These are all the tiles between Las Vegas and San Francisco. Because the system is very smart and keeps a persistent cache, we can actually look for differences and only download where there have been changes, again, to save bandwidth and to gain speed. The less you have to download, the faster you can actually start driving, which is happening here. We've developed this. It's called AutoStream. When we started, people said, "That's not possible, right? An HD map is much bigger than a navigation map, and how on earth can you stream that?
We can't be dependent for safety critical function on a cellular connection. We managed to address all of those problems and created AutoStream, which is going to be in a production vehicle next year. That's AutoStream. That's the centimeter level accuracy, millions of sources in real time. Then we move to the final bit I promised, which was euro numbers. How big is this market? What I'm showing here is multiple opportunities. You see vertically, you see private vehicle automation. That's the level 2 and 3, and a little bit of level 4 in the future for private vehicles. Vehicles that all of us would buy. Then there's robotaxi or shared vehicle automation. Different market, different business models, partially different technologies, but very similar maps.
There's one example of Other, there's actually a larger category called Other, but this is one example. Asset management for government, right? If you have all this very detailed data, you can actually also add, and find things like lantern poles in the map. You can find electricity cabinets, all that stuff. That is of interest for municipalities and other governments. You see 2020, 2025, 2030. If we look at private vehicle automation, in our own market model, we see that the market in 2020 is roughly EUR 0.1 billion, right? This is what the model gives us. I'll explain a bit about that in a minute. 2025, by 2025, it has grown to EUR 0.8 billion. By 2030, it has grown to EUR 3 billion. There's an enormous growth in the private vehicle automation market. If we look at robotaxis, we think it looks like this. Yeah.
Similar size, EUR 0.1 billion in 2020. In 2025, not yet very big, EUR 0.3 billion. A lot smaller than private vehicle automation. It'll go in 2030 when all of those difficult technology nuts have been cracked and there has been time for scaling, it'll be around EUR 5 billion. Government, asset management, smaller markets, EUR few hundred million. Now, everything we do, and it's important to understand, and everything we project, is only based on private vehicle automation. Yeah. That's what we focus on. That's the business where we are leading, that's where we want to win, and that's what's driving all of our decisions. Robotaxis will come, but it's upside in all of our projections. We go a little bit deeper into private vehicle automation. This is part of our market model. Let me explain this to you. You see horizontally the years.
On the left axis, you see tens of millions or 20 million, 40 million cars that have a particular function. On the right-hand side, you see % with attach rates. How many of these cars will have an ADAS map or an HD map? Those are the two lines. The colors in the graph, light blue, medium blue, and dark blue are level 0 and 1 combined, level 2 and 3 combined, level 4 and 5 combined. What you see is that there's already on the road next year, more than 60 million vehicles with level 0 and 1 functions. Often they don't use a map, right? The attach rate of maps is around 20% of ADAS map.
As we see higher and higher automation coming in at a level 2, 3 functions, we see also the attach rates of maps growing, with HD map growing faster than the ADAS map. Ultimately by 2030, we expect to see, or actually the analysts that we've based this model on, expect to see level 4 or 5 private vehicles again. These are not the robotaxis, being meaningful as well, being a meaningful percentage of the total amount of vehicles. You basically also see that nearly all vehicles will have some form of automation. Yeah. We sell in the world about 100 million cars a year. That will grow a bit, so that has gone in to those market estimates that I just gave, right? EUR 0.8 billion in 2025 for private vehicles is based on these take rates and these volumes.
What does that mean if we split it out? This is the ADAS map market. You see, again, the same time scale. You see the total market value, and you see that by 2030, this is around EUR half a billion. That's what you see all the way in the right-hand side. Not all of that is addressable. Part of it is in China, which we can't address, and that's the dark red part. This grows about 20% per year, we expect, yeah, in ADAS map market value. We do the same for HD map, it looks like this. Yeah. We actually see it starting from a small base. We see a 60% growth over the next decade, 60% CAGR over the next decade, with a very large part, nearly EUR 2 billion by 2030, addressable for us. Yeah.
This is what we work for every day, and this is what we try to capture as much of as possible. Where do we stand today? Well, the numbers have been mentioned before today. If you look what got us here, it's our competitive advantage. Yeah. We have experience with dealing with automotive customers, customer experience, customer intimacy. Through our automotive sales force, everything that Antoine is doing. Through our cumulative domain knowledge, our global presence, right? We are everywhere, both in mapmaking as well as in sales. If we're not there, we travel. Our independence, the Switzerland of maps, as Harold said in his introduction. That's one element of our competitive advantage in this space. We have leading technology. A lot has been said about that as well. Our transactional mapmaking applies equally well to our ADAS map and our HD map.
That's delivering the shortest cycle times and we've built up a lot of technology over the years in artificial intelligence, computer vision, and what have you, right? That's also all in all, giving us a leading technology position. Finally, we have a very complete, the most complete product portfolio in this space. You've seen Kees talk about integration with navigation. We have the SD map. You need an SD map to be able to make an ADAS map, right? An ADAS map is a road-level map, and we attach attributes to that. You need an SD map. If you don't have that, you can't play. We have AutoStream, I mentioned that. We have services, they have been mentioned, and we actually can help to integrate all of those products into the vehicle, right? By working with partners where we do pre-integration, we have our own integration services.
All this together gives us a very strong competitive position. Nobody has the same set. We're better than HERE in many of these things. We're better than Google in many of these things. That is translating into market share, right? More than 1 million vehicles, level 1 and 2, mostly level 2 actually, on the road today as we speak, powered by our ADAS map. We're learning from that every day. We get feedback from the OEMs. They tell us what they like about the map and what they would like to be improved. We work with that. We bring that in. In HD map, we're the current leader. We're the market share leader, the market leader also in technology.
This is expressed as a market share of 60% out of all rewarded deals, sort of between call it a year ago, 18 months ago, and two years from now. Every single OEM will have made their HD map selection decision. At the point in time where we are now, we are at 60% market share based on awarded deals. That doesn't come out of nothing. It comes out of our competitive advantage that I just described and out of working with all of these OEMs for a long time. Erik also mentioned how long we've already been working with Zenuity, even before Zenuity existed. We've been working with four out of the top five global OEMs when they were developing their test vehicles and their R&D vehicles. We gave them map samples. That has translated for some of them into wins. We're the only company with a self-driving stack.
The car is outside. We test our own map. We eat our own dog food, as they like to say on the West Coast. We had many firsts in HD mapping. We were the first to file patents in 2009, the first to provide samples in 2013, the first to provide a country in 2015, and the first to cover the highways of the world, at least where those vehicles will drive, in 2017. Yeah. With that, I would like to thank you for your attention, and I think that's time for questions now.
Quick question, of course, on your market expectation for the next few years, especially in automotive. I wanted to ask you maybe, Antoine, first, in the next few years, excluding ADAS, how do you see the content per car evolving? Meaning, do you see any kind of pressure on pricing? You have a lot of services, more and more attached to it, like you described EV. Is it something that you can increase basically the value you get per contract, or is it just to offset basically underlying decline of the price? It's first question.
Price pressure is not new, right? That continues. I think it's a bit of both. If you look at what has happened in navigation maps so far, yes, you're right. Basically, integrating more content or more coverage compensates for the annual price decrease. I think if you look at what we're doing now, including increasing the size of the services package, it's more than that. There's added value on top of what we currently discuss.
Okay, great. On the HD map, I wanted to ask the same question because separate, it looks like the OEMs are facing a lot of challenges to some extent with electrification of the car, with ADAS, and you see the bill of materials increasing everywhere, basically, almost. I was just wondering, given the HD map cost, possibly, we are talking about an increase of content significance. That's what you are publicly saying. Do you think for penetration to increase materially like you're describing, you will need to decrease the bill of materials as well materially? My question, in your market assumptions, do you expect the value to come down, the ASP to come down for the penetration to go up?
In the market model that's behind these numbers, there is a price decline at similar features assumed. Yes, absolutely. Because that's indeed, as Antoine is saying, that's the nature. There is price pressure in the automotive industry. As features increase, of course, we hope to compensate some of that price decline. I'd also like to say a word about what you said about, is it expensive? Yes, electrification and ADAS and all of that, it's expensive. The part of maps is not necessarily the biggest expense, right? You're going to put sensors of EUR 100, EUR 200, several of them, you're going to put in a vehicle. The cost of a map, which scales much better than a sensor, which is hardware, in the whole system, is not an enormous amount, an enormous percentage in the whole ADAS or autonomous driving system.
Last one from me is if we are looking at your forecasts for EUR 800 million of value by 2025 for HD map.
If we look at your market share above, I mean, 60% that you are mentioning, let's take 50%. Like EUR 400 million potential revenues from HD map we can expect within five years, basically also. Is it the right way to look at it?
Your calculation is-
Correct, yeah. Yeah. Revenues, we're not talking about orders intake, it's just revenues.
I didn't get the last-
It's revenues, not order intake we are talking about.
No, that model is based on revenues.
Okay.
You get subscription revenues, right? It's different. It's more like a traffic service where you get annual income from the map.
Okay.
I'm not going to comment on what IFRS will do to all of that, but I'm sure it will do something.
Thank you.
Can you maybe give a bit of an update of all the partnerships? In the past, you have announced partnership with NVIDIA, for example, or with Bosch on this level, or even with VW. How does it look today? Who's really driving this autonomous driving trend, and how do you partner with them? What's the status there?
Status of all the partnerships. Yeah. We've made a lot of announcements on partnerships. The key thing is that you want to pre-integrate, basically, especially if it's other automotive suppliers, you pre-integrate. Zenuity is a great example where we work through all of the challenges that you face when you start to collect data from a vehicle. Volvo is not necessarily needed for that. By the time Volvo, as a potential customer, would come into the picture, we can say together, "This part of the puzzle has been solved, and it will make your vehicle safer." How are all these partners progressing? Some more rapid than others. I visit most of them, and we try to continue on that path of mutual learning and do that pre-integration.
For example, the NVIDIA partnership, for a while, it was a lot of noise. A lot of noise coming from NVIDIA working on autonomous vehicles. It's a little bit more quiet now on that front. Is that just perception from us from the outside, but internally you're working a lot on a lot of products with them?
I'd say in general, this is not TomTom specific. If you would look at the amount of press releases that were done, you would see sort of a peak in 2018. Now everybody is sort of hitting reality. Cars have to hit the road. Business is one. Let's get our heads down and deliver, and learn from that. I think it's correct that you're hearing less, but a lot of work is still being done behind the screens to realize all the promises that we made in those press releases.
Final, second question is more on the revenue model. The way you talk about it, and probably that you will have a connected car all the time and a continuous update, it really feels more like a software as a service kind of product, which should have monthly or yearly subscriptions based on the users. If I'm correct, the OEMs are not looking at it from that side. They're looking more from a license product. Is that changing, or do you see that moving towards that's more like a SaaS product?
Well, it's been an ongoing discussion for a while already with the connected services. Connected services have not always been on just one license and done. The more we move, just as you mentioned, the more we move towards navigation as a service altogether, the more we deliver regularly or continuous map updates, even more when we come to software updates. It is going to be a navigation as a service model, and therefore, more a subscription model than just a one-time license fee. Those discussions are taking place. It's a change, but I wouldn't say there's a strong resistance from OEMs to move towards that direction. It's a change.
Yeah. I have a question. Yeah, Nigel. Nigel with the Capital Coast. Question for Kees. Maybe, I think you call yourself the chef with the recipe in one of your slides. To what extent are the OEMs allowing you to be that party? I mean, why wouldn't they want to be that?
I think across the board, we see a very spread level of maturity at OEMs with regard to developing that user experience. It's difficult to do software design in the car, and it's very different from how, for example, design organizations at car companies work. I think, across the board, they still spend most of their time on the brand and identity of the car, of the interior, of the hardware, user experience of the car. Software is, for them, almost alien. For some. For the less developed OEMs. If you look at your high-level quality premium companies, it's different. They have a very good understanding of what they want, and they are more likely to be customer of our ingredients. Components that we deliver, like maps or services or online routing and search, for example.
There's also a large class of potential OEMs that would love to have.
A solution delivered to them on a silver plate that they can customize to their liking and to their needs, and have all the hassle, I would say, around software updating, software management, having an integrated experience left to a potential third party. Yeah, we definitely believe there's opportunity there in that space.
Maybe a follow on then. Is that exactly what Google providing to these guys? Like a one-stop shop, and they provide everything?
Yeah, it's difficult to understand really what Google is doing, but if we learn from the past, it seems, for example, customization options are very limited. It's very much, "This is our solution, take it or leave it." That's what we've seen in the mobile space, right? If you buy an HTC phone or a Huawei phone or any other Android phone, it's a Google experience, and it's not an HTC experience. I think that is something that the OEMs are really trying to avoid. They want to create an experience that is a brand experience for their customers. You need a certain level of customization of that software use experience that is fitting those brand needs. There is opportunity next to a Google-only solution.
The final question, and then over to you.
Okay.
Can you provide, or is that sort of catered for in the current R&D budgets, that you'll be able to connect all these apps within the car to each other and make your own, well, you provide the recipe for the OEMs, or is that not the right way to look at it?
That's a case-by-case basis, right? One of Google's strong points is, for example, having an app store. In theory, any application builder can create an experience in the car. We also know what people do in the car is a very limited set of use cases. It's very different from how you're using, for example, your phone in many other contexts. When people are driving, and they pay attention to the road, it's different once you're in an automated driving use case, people just flip over their laptop or take their tablet and do whatever they want. As long as you're in a driving situation where you need to require attendance, what people tend to do in the car to make best use of their time is to communicate, to listen to entertainment, and to find their way, right, to navigate.
Especially communication, whether it's making calls or conference calls or getting ready for your day at work, those are the typical kind of communication productivity use cases. The other class is just listening to the music I want to listen to in the car, or that radio channel that I'm used to listening to, or continue to listening what I was doing, listening on my phone when I jump into the car. That set of use cases is rather limited, right? You don't need to boil the ocean in order to make those use cases actually work in the car. It's providing to a couple of key services, having a great interactivity with your phone so that your car is a fantastic remote control to whatever is running or playing on your phone.
Nailing a few of those critical use cases essential. Having an enormous thousand-app app store to watch the news in the car or to look at your stock ticker, we don't believe that is a requirement in order to deliver a satisfactory, delightful use experience for what people actually expect to do in the car. It's a bit back to that iPhone use case, right? The first iPhone did a lot less than what a Nokia phone did at the time, a lot less. The things it did well, it did totally frictionless, and brought user delight and brought user satisfaction. I think there is also that opportunity in the car. This is not about enabling each and every use case, but this is about doing a couple use cases really, really well in a delightful use experience. That's kind of what we're aiming at.
Marc, go ahead.
Marc Hesselink, ING. The EUR 3 billion you mentioned as a sort of potential revenue pool in 2025, and also with a CAGR of 16%-20%, how does that fit in with the CAGR that Taco gave in his presentation towards 2030, and the order book that is attached to that? How do these numbers link in with each other? Should I see this separately from that one? Is it part of the order book that we should see, say, 10 years from now? How does that interlink?
Well, obviously what we're doing in ADAS and HD is going to be part of the ambition that Taco stated for 2030.
Given that the growth rates are much higher and the potential quite big, 15 versus 60%, how does that-
The growth rate of HD comes from a very small base. It's a bit misleading to use that 60% CAGR and compare that to the 15%, which is from a much larger base, which is our current revenue.
What would you expect then to be the proportion of HD to be in that number by then? Is that possible to give? That's a bit difficult?
Let's not go in that detail, Marc.
Okay, good. Maybe over to Zenuity, because you probably were there when the decision was taken from Volvo to make a U-turn and turn to Google. Can you explain perhaps what the level of reasoning was behind that? Maybe follow up, a few years down the road, can we expect another U-turn? Are things still going in the right direction, or would it actually be more TomTom in the future?
Okay.
Anything you can comment on that?
I was there. I cannot comment, as I'm not a Volvo employee anymore. With Zenuity, a little bit as Willem was hinting, we are the active safety self-driving car software supplier to Volvo Cars. We're working intensively on that stack together with TomTom. The actual sourcing decision to source a TomTom map or to somebody else, that's a Volvo Cars decision. I cannot comment on that then.
As you would have expected.
Of course. Thank you. Wim Gille from ABN. You mentioned during the presentation that so far you've won about 60% of all the HD map awards. Who's winning the other 40%? Is that just HERE or are there also other parties involved in that 40%? Maybe also a bit on we've always had a bit of a bumpy ride in terms of order intake because the level of RFPs is not the same level every year. It's very difficult for the outside community to see what the level of RFPs is. Can you give us a bit of glimpse into the future in terms of what the level of RFPs has been in 2019 versus 2018? Also maybe moving towards 2020 and 2021. Is that number going up or should we expect more soft patches there, as an industry as a whole? Also a question for Zenuity.
The choice has been made to work very extensively with TomTom. What was the reason, the final reason to work with TomTom, is TomTom your only partner, your exclusive partner? What were the key tick in the box that made you choose TomTom for such an important partnership?
Let me first answer the first two questions. On the market share that Willem provided, that's a volume-based market share. It's not that there were five awards and we won three of them and that makes 60%. Based on the car volumes, we've won 60%. On the RFQs, we see similar trend as we've seen last year in 2018. It isn't that big as it was in 2017. With introducing the automotive order or the backlog, we think it's also more fair KPI to see future revenue. We continuously see reassessment of already running deals and extension of running deals, et cetera, that you can't report in the order intake, but you can report it in the backlog. The best future indicator of success for automotive is the backlog and no longer the order intake.
For the last question, do you want to take that?
Oh, yeah.
All right. You first or?
No, you go.
Okay. I can answer on why we decided to work with TomTom already in the early days. Most of it's willing to explore. Developing HD maps that fit localization, that fit safety critical ADAS, safety critical automotive driving, is a journey, and we didn't have all the answers at the beginning. TomTom had a willingness to explore and had HD maps available, can do really fast map making, so that if we do the probe sourcing mechanisms, detect new features, send it to the cloud, update the map, this entire machinery, and there we felt that TomTom was the right partner for us, and we still think that. Is it the only map maker we're working with? No. It's the only map maker where we really explore the new grounds and where we do the core development with.
Very good. You really can't comment on who's winning the other 40% of the volume in HD map making. Is it just HERE or.