Good afternoon. Why don't we go ahead and get started? So I'm I'm Robloff, our CFO. I want to say thank you very much for joining us at our at our Analyst Day. Delighted to be here.
We have certain members of our management team, including, Luke Sarthon, our CEO here as well, and we have a presentation that's also available on our website. Go through our presentation, and then it'll be available to download after our presentation is over. Just in terms of the order of operations, after my brief introduction, Luke will spend a lot of time more talking about our company, and I'll come back, shed some more light on our financials. We'll have our management team come up and join us for any questions that you may have about us or our company. I want to start off just with our obligatory safe harbor statement, Obviously, we'll be making some forward looking statements today, not just about this quarter, but this year and next year as well.
I'd encourage you to read our filings that are there with the SEC. I think they provide a lot of useful, content about our company, certainly more than we can provide just in, in a short conversation. So with that, what I'll do is I'll turn it back to Luke to kick off.
Good afternoon, everyone, and welcome to New York City. Welcome to this wonderful building and welcome to our 2019, Analyst Day. The way we are going to structure the presentation today. I'm going to introduce the management team. I'm going to talk about the end market that we are targeting as a company and why they're relevant to us.
We'll talk about how we're going to describe our business going forward We spent a lot of time over the last year trying to simplify our business in the decisions we've made. We also want to simplify the way we present our business So that's easier for you guys to track what we do. I'll talk a bit about, our product offering and it's relevant to the target markets we have. And finally, something that we think is very important to us, we're going to talk about our R and D activities and how they fuel our patent portfolio because we had a lot of questions from that topic, and I think it's important that we cover that. So with this, I'll start with this very simple slide That kind of captures who we are.
We are RAM Bus. It means that we make memory and high speed interface RAM and we move data, we move and we secure data. And that's how we want to position the company. We want to be understandable by the market, and that's how we are being understood by the market. So I start by introducing the management team, some of the people that are here, to see there's been a lot of changes over the last year.
So let me take a few minutes to introduce the people. So if you look at the top role on the right side, This is Keith Rogers. Keith is there. Keith has been with us many, many years. He's coming from CRI, the embedded security company that we bought years back, kit has been, front and center in our patent licensing negotiation and is also in charge of our M and A and corporate development activities.
So we have an opportunity to meet with Skip today. Next to Keith is Jerome, our CMO is not here today. He's in Europe. Geoff Moore, next to him, is our head of operation. Jeff Moore is coming from Fairchild.
He's the one who has been behind the infrastructure we put in place become a chip company, very good experience in supply chain management in particular. The next one on next to Jeff is our friend Rahul, who's going to talk after me. Next is Jay Kim, our General Counsel. And finally, have the pleasure of having Sean Phan with us. Sean is there.
Sean just joined us 3 weeks ago from IDT where it was running for business units. They are including the buffer business unit. The way we've organized the team is that all of our product activities are going to now report to Sean. So you have Gary Brunner here. Gary experience has been with us for a long time, but before that, it was with IBM Research.
And Gary is looking at all the innovations that we have to go through in the memory space. And he reports to Sean. Krishna is our head of a design center in India Neiraj is also, Paraguay is also someone we hired from the industry. He was in charge of the security business at NXP. And he's in charge of our security business now.
Herman Duer is in charge of our IP business. His background is with Xilinx and Intel. And finally, Chancein Lee joined us years back now from Inphi and his background was in the buffer chip. Business. So as you can see, we simplify the structure of the management team, we're focusing the management team, and we brought talents from the industry to help us grow the company.
At the beginning of the year, we laid out a few priorities for the business and of you are very familiar with those. The first one was to refocus our product portfolio and R and D around our core strength in Semiconductors, in particular, memory interfaces, high speed interfaces and embedded security. We also put a lot of effort in optimizing our operational efficiency by leveraging the synergies that we have between our different businesses and across our customer base. And finally, we leverage our ability to generate cash from operations to amplify our market and technology growth. And you see some of the things we've done over the past few months.
A brief summary of what has happened over the last year. If you look at the top line, this has to do with the strategy that we laid out a few months back. In the end of Q3 of last year, we announced that we would be looking for strategic options for our payment and ticketing business. Remember, payment and ticketing business was very successful but it plays in a set of markets and in an environment that is far afield to our core. So to allow them to continue to grow and to allow us to continue to focus on our strengths, we decided to divest that business.
In Q2 of this year, we announced that Visa would acquire our payment and ticketing business. And we are in the final stage of that process as we speak. The other thing that we did on the strategic side is very recently, in Q3, we announced the acquisition of Northwest Logic, a provider of memory and Sirdes Conforters and Nippy Conforters that really complement agreement to acquire the embedded security business of varying metrics, originally known as Insight Secure. That strengthened our position in the embedded security space. So on this strategy, we made a few moves 2 of them just happened this quarter to strengthen our position on our core business.
On the product side, we've made a few announcements over the year about the progress we've made with our crypto manager. The first one was with Micron that the announcement we made in Q4 of last year. We've expanded our crypto manager business with other customers that we don't name in AI and in consumer space and that happened in Q2 of this year. We've made some progress in the IP core business by announcing the release of our 32 gig and 112 gig 30s. This 30s is extremely important in markets like 5G, for example, or the next generation of network infrastructure.
On the licensing business, the last quarter of last year was very busy. We had a lot of critical renewals. Our Friend Kits was very busy doing this. We renewed our licensing businesses deals with Broadcom, Nvidia, and Fysen. And last but not least, as we do all of these, as we redefine the company and refocus the company, we continue to generate a lot of cash from our operations.
Last quarter, we generated $38,700,000 of cash while redoing all of these activities. So that summarizes what has happened over the last year. As I said, we had a sequence of announcements over the last few months. The sale to Visa of our panel and ticketing business, going through its final stages of completion. We acquired and I'll say a few more words about this later.
We acquired the business of Northwest Logic in the controller space. And we made an announcement to require the embedded security business from very metrics, known as Insight Secure. This slide is a good summary of what we want to be and what space we are playing in. And if you look at the bottom of that slide, what you see is you see a list of technology suppliers. And If we want to define who we are in terms of what neighborhood we play in, then that would be there.
Our typically our competitors in the IP core space, whether it's 30s memory or security, are companies like Cadence and Synopsys or Arm. Typically our competitors on the cheap side are real competitors today on the cheap side are IDT and montage. So this gives you an idea of what space, what markets we're trying to address. We have a very strong relationship with the foundries, Samsung, TSMC And Global Foundries for a very particular reason. This particular reason is that when you develop memory interface is 5 or service interface is 5.
These files are based on analog technology pendant on the technology node that is being used. So we need to have a very strong relationship with the foundries every time we develop a 5. And that gave us that strong relationship with those companies. If you go to the next level, This is a shortlist, a very shortlist of our customers. You see names like the memory vendors or Qualcomm or Intel.
There's a long list customers. But I think what's important here is what you can see is a lot of these customers are licensees. They were licensees of our company's years back. And over time, what we've done is we've reinvested the cash that we get from these companies to develop solutions and technologies that they need. So our relationship with those customers is much broader than the licensing relationship.
It's a product technology and licensing relationship. The next layer talks about the OEMs. The OEMs are extremely important for us because they define the cadence of introduction of new platforms in the market. That's extremely important for us, especially in the buffer chip business because they define when the next generation of products is going deployed in the market. It's also very important because they help us define what we need to have in our products.
And anticipate those needs when we build those products. And finally, the cloud providers are becoming increasingly important for us for two reasons: one, is cloud providers are starting market, in particular, in particularly, around AI. And they also are in high demand of embedded security. Because they play in the cloud. So we have a growing relationship with those companies in the cloud.
And on the top line, what you see is that, originally, Rambus was known to play mostly in the data center space because we were focusing on memory. But there's a renaissance of interest for memory technologies and embedded security technologies in a lot of different markets. And I'm going to describe a few of them as we move forward. Here what we see is the markets we are playing we're known for playing in a data center and networking. We have great traction in automotive and internet of things, AI, edge computing and governance.
I'm going to click down on a few of them to explain why these markets are very relevant to us and why we decided to refocus the company on memory interfaces, high speed interfaces and embedded security. So let's start with the data center networking. Everyone knows that chart that shows the explosion of data, that we are all experienced saying. And it's mind boggling. The scale on the left side is in Zettabytes, and people just have trouble understanding what Zettabyte is, I try to find myself.
So Zettabyte is a 1,000,000,000 gigabyte, So and you see the market size in the sort of range of 20 or 40 Zitter bytes, billions of gigabytes. It's very difficult to fathom what it means. But I think the way to look at this is, to look at our life every day with our mobile phones. People don't know, they are probably in the clouds, 4, 5,000,000,000, 5,000,000,000,000 pictures stored every year. Every minute, there's maybe $120,000,000, $130,000,000 emails being sent, $16,000,000 text.
It is really mind bubbling. They're 1,500,000,000 people active at every moment on Facebook. So that creates an amount of data that is really, really difficult to deal with. And I think the expectation is for all of us of, real time access to that data. So the way that chart translates to us is that as you see this exponential growth of data, if you want to have real time access to that data, it means that we will have to have this exponential increase in the speed of the solutions we provide.
If you want to have access to something that is 10 times bigger in the same time, then you have to go 10 times faster. And that's what this company has been doing for the last 30 years. This company has been developing high speed interfaces that go faster and faster to access larger amounts of data faster. And that's why that chart is very relevant to us. The data center market is a large market for us.
We estimate the time to be $800,000,000 with a lot of design starts with a growth that is a modest growth of 8% but that's the bread and butter of what we do. What you see in the center of that slide is the types of products that we have to address that market. And in the case of Rambus, all of our product portfolio have a role to play in the data center space. Chips, buffer chips, memory files, memory controllers, 35s, service controllers, security cores and security provisioning. Every one of those products has a role to play in the data center space.
So as I said, this is the largest target market for us with good growth and a very nice focus ai is one of these transformational aspects of the market that we can benefit from. AI is based on the concept of learning and and then making inference about these learning, training and making inferences. To train a AI system, you have to train the AI system with tons and tons of data. You have to give, for example, the zillions of voice samples for an AI system to train to recognize your voice. So AI is actually in very, very high demand of data.
And data bandwidth. The higher speed and higher bandwidth type of solutions that we offer to the market go to AI type of applications. And there's a very interesting interdependence between AI and the amount of data we have to deal with. We, as people, cannot make sense of that amount of data that is out there without AI. We need AI system to make sense of data.
At the same time, AI systems can deliver good quality results if they can train on a lot of data. So it's right in the middle of what we do, and that's why we have all of these new design wins in that space. And if we do the same analysis as we did for data centers, you can see that the AI market TAM is smaller to $125,000,000 according to us. And, that the number of designs is quite nice, but the growth of that market is much nicer than the growth that we have in the data center. So we have that view of a very large market data center, $800,000,000 TAM with a growth of about 8% and an AI market, that is growing much faster, smaller market today, but growing at a much faster rate.
And what you can see here as well is that the type of products that we have elected to play in are very relevant to the AI space. So in the AI space, we sell memory files, memory controllers, 35, service controllers, security course. Security provisioning. The only thing we don't sell into AI today is buffer chip. Let me move to IoT.
So IoT is a new source of data. What we talked about earlier was what we as human need in terms of data. We talked about pictures. We talked about Facebook. We talked all the things that matter to us.
Now with this emergence of connected devices, we have a whole new set of data, which is the data we read from connected devices. And here, one of the concerns of the whole industry is how do we make sure that the data we read is secure and is authentic. So that we can use that data for whatever service we want to base on that data. So the IoT market is still smaller for us, in terms of 10,000,000 dollars, $60,000,000, but there's a very large number of design starts and a nice growth as well. And the way we play in the IoT market is memory controllers, service controllers, security course and security provisioning.
Again, a very nice fit of what we do with what these markets require. So if we lump all of these together with the other segments that I did not describe in more detail, What we see is we have a sum of $625,000,000, meaning that if we stop developing any products and stop acquiring, this is the market that we're looking at today, $625,000,000, for our product offering. That excludes the licensing business. Now if we go to the next circle, the SAM is $1,400,000,000. And we make dense into the SAM when we do acquisitions, for example, when we acquired Northwest Logic or when we acquire Verimetrix, we acquire technologies that are very close to our core and we can address that time of $1,400,000,000.
The time of $2,000,000,000 takes into account the fact that some of our customers develop their technologies themselves in house. And if they were to have external companies develop those technologies, then the time would be $2,000,000,000 for us. So in a nutshell, this is what we do and why we do it. Originally, we were a company in the memory space, mostly addressing the data center space. We see a renaissance of memory architectures and the need for embedded security due to the emergence of AI, IoT, automotive, and other types of markets.
So it's a kind of a rebirth for Rambus, and it was important for us to double down on the technologies that we master. As we address these shifts in the market. Now as I said earlier, we wanted to simplify the way we present ourselves. There was a lot of effort for people who follow us. There was a lot of effort to simplify ourselves internally.
To be clear about what we want to be, what we want to develop, what markets we want to go after, and that's what we just described. I think it's important for us also try to be clearer and more understandable to the external world, including the financial community. So The way, you know us is this way. We had the memory and interface division and the security division, and we were addressing each one of those divisions with the 3 ways of going to market architecture license, which is basically patent licensing, interface IP, which is basically IP cores. And chips.
In the case of security, we had the same thing. We have patent licensing, we have IP cores, and we have payment and ticketing. Which was part of our business. So as we simplified our business, we got rid of payment and ticketing. As I said, payment and ticketing was growing nicely but in a very different market, it was very distracting and it was preventing us from investing in what we're good at.
So we are divested payment and ticketing. The next thing that we're doing is we're simplifying the way we look at our business. On one side, we have our patent licensing business, which embraces everything we do in mid and RSD in security and memory space. We have a IP business that is very similar in terms of how we go to market between our security IP and our interface IP, mostly because our customers are semiconductor companies. And the type of contracts we have are very, very similar with those companies.
And on the right side, we have a cheap engagement, which is a standard cheap engagement. So we think that the way to present our business as delivering technology in the embedded security space in the high speed memory interface in the form of architecture license, IP, and products is a simpler way of describing our business. So the new runbus is the way to look at this. We have an architectural license business, We have a silicon IP business that embraces everything we do in security and IP course, and we have a chip business. And we deliver those technologies to the market through these three channels.
And hopefully, that will simplify the way we describe our business. And I think Rahul is going to talk in more detail about how financially we're going to describe this to allow you to understand us better. What I'm going to do now is give a brief introduction of the product offering that we have in place to address all of these target segments. The first one is the memory buffer chip. A lot of you are very familiar with this, but the memory buffers are expanding the capacity for data centers as it sits between the memory the DRAM memory banks and the processors.
I think what's important to know for us is that today, the buffer chip business is playing in the generation of DDR4. But DDR4 has several flavors. Every time a processor vendor introduces a new processor or every time a memory vendor introduces a new version of their memory or DRAM memory, there is an opportunity for us to gain share by winning designs with our customers, our direct customers, those memory vendors and their customers. That we looked at earlier. And the growth that you see in our buffer chip business is stemming directly from our ability to win designs every time there was a change in that market, either a new processor or a new memory being introduced in the market.
We've seen very nice growth in that business, as we said earlier, more than 40% growth year over year in that business. That business is going to be anything between $50,000,000 $70,000,000 this year, and we'll continue to grow. So very, very nice business. And all of this is coming from the DDR4 space. We have a nice niche business are in the non volatile buffer chip space.
I'm saying niche just because the end market is not because we have a small share. We have this big share of a niche. And that's one of the technologies where we adapt our buffer chip to hybrid systems where we want to mix a flash memory with DRAM. So I just wanted to mention that because this is a nice growth us, this is also a very interesting system work that we do with some of our customers. And on the right side, as you know, we were the first ones to introduce DB chips and RCD chips for the DDR5 market.
We expect DDR5 to sample next year. No volume not before 2021 at the year earlier. So there's a lot of business to be done in DDR4 before we go to DDR5. But the very fact that we were first in DDR5 in sampling is very comforting for us because history showed that when your first market and engage with the ecosystem earlier, then typically you drive a large market share. This was the case for Inphi in the DDR3 space.
This was the case in DDR4 with ADT. And we expect that this will continue to be the case for us. With DDR5. Let me explain what we do in Silicon IP. IP course and why it is relevant to the markets that we talked about.
If you look at a card in a server here and a processor, you look you can see all the areas where the IP cores are relevant. So On the left side, you have memory interconnects, all types of memories that you can address with this processor card. HBM2 GDPR, DDR, LPDDR, all of these memory interfaces are offered, as by Rambus, and we've complemented this with the memory controllers that go with the acquisition of Northwest Logic. Every card on the backplane potentially connects to long reach 56 gig, 112 gig SerDes. Same thing, these are developments that we have internally.
You have chip to chip interconnects, PCI, 3, 4, 5, 12 gig medium range, 28 gig, again, all sorts of 30s that we develop in house. And when we bought Northwest Logic, we complement those studies with the controller that goes with it. Some people are talking about chiplets, which is an interface chip that sit between different chips that build large systems. So our strategy here was to focus on high speed interfaces because this is hard to do the barriers to entry are very, very high for people who are trying to do that. And to focus on increasing the ASP that we offer to our customers by buying, for example, the controllers from Northwest Logic.
Because for every design that we have for 5, there's a potential design for a controller. The side effect of that as I explained earlier, when you design a PHY, you are dependent on the technology node and the foundry that you work on But when you design a controller, you're not. So what it means is that our Northwest Logic friends or our Northwest Logic employees now I've had a very interesting reach to the market because they were not dependent on technologies. And this synthesized about IP. This slide shows the complementarity of Rambus and, Northwest Logic So we originally developed Fis for DDR4 GDDR6, HBM2, PCIE we didn't develop, files for BP.
Northwest Logic had almost a mirror image of what we do in the controller space. And this was a very natural move for us to do that because we increased content and our end customers. We also, improved time to market because the inter operability of the 5 with the controller is much easier when you're both part of the same company. So you can actually work on systems that work together much faster and you can get to market much faster. And at the bottom, you see the markets that we're addressing, which match what we talked about earlier in our presentation.
So let me switch to embedded security because we made an announcement, recently. Embedded securities about protecting Semiconductors and their secrets from design through manufacturing and deployments. And as I explained, earlier to some of you, the 2 parts of what we offer, their secure IP course that are embedded in silicon, where typically you store secret keys. And when you access the secret keys with a mirror key, you have the possibility of enabling or stopping or changing things in a secure way. And we provide the infrastructure but it allows you to have provision those keys to these embedded embedded silicon.
This is very, very important. People are becoming more and more aware of the need for security and the risk of having all of these IoT devices, for example, being insecure. I was reading recently some articles about what happened, I think, in 2017 in Dallas, where 156 CRMs started to ring in the city for 2 hours because they were hacked. And it floated the nineeleven numbers. People in the middle of the people in the middle of the night were panicking about what was going on.
It is very important that every time you use a device that is not connected to a human, that you make sure that this device is secured, that is authentic and that the data that you sent from that device, or is this an to this device is data that you can trust. So that's what we do. What you see on the right side is the product offering that we currently have before the acquisition of Veri Metrics. But the next slide shows how we complement each other with the Silicon IP of Veri Metrics. As you can see, we both provide secure silicon IP.
We both have your different segments But one interesting aspects of what, we did with, very metrics is that originally, Rambus was about securing data at rest. You're securing the data that sits in a piece of protected silicone in in the chip. Very metrics add to this, the ability to secure data in motion And there are markets, for example, in the networking space where you need to encrypt the data before it exists the chat. For security reasons. And very metrics brings that capability that we didn't have in house.
So it's very complimentary from that standpoint. They also bring a secure software protocols, and they also have their own secure provisioning system. With the acquisition of Verimetrix, Number 1, very metrics can focus on their own business, which is higher up in the software stacks. So it was good for them. And for us, it reinforces our position as a leader in embedded security systems for the silicon market.
Semiconductor market. So now that we have an overview of the current product offering, And as we talked earlier about our effort to refocus the company, I just want to spend a few words about what we do in R&D. You remember one of the aspects for this refocusing of the company was the moves we made through M And A, the divestiture of our payment and ticketing business to Visa, the acquisition of Northwest Logic in the controller space, the acquisition of Verimetrix in the embedded security space is all part of refocusing the company, but we've had some efforts in refocusing the company internally as well. We've cut programs in the R and D space that did not had customer traction. And we've redirected those R and D dollars into areas that make sense for our short and long term future, in particular, to our licensing programs.
Licensing programs is a very important part of our business. It generates cash, it's stable, it's long term. And it's very important for us that we continue to invest in that program. So there are 3 in the memory space, there are three areas where we spend our R and D dollars. The first one is extending the performance of DRAM for future generations.
And basically what it means is that, what do we think are going to be the techniques we acquired when we go beyond DDR5 into DDR6. When we go beyond GDDR6 into GDDR7, and so on and so forth. So we have a group of people who look at techniques that we think are going to be relevant to this next generation memory that are not in the market yet. We're just inventing, we're just producing invention We're just patenting what we can patent. And all of these activities under Gary's group here.
The second area we invest in in R&D is the next generation hybrid memory subsystems. And this is the techniques that we think are going to be relevant when we want to mix on a memory bus, some volatile memory like DRAM and some non volatile memory like Flash. And what system requirements are going to be. And again, we're trying to invent techniques that we think are going to be relevant when those technologies are going to be widespread in the market. And finally, the 3rd area we're investing in is, the application specific memory solutions, in particular for artificial intelligence.
As we said earlier, The current memory subsystem architectures that we use for data centers are not optimal for new applications like AI. AI requires a lot of data for training in particular. So we need to find ways of increasing bandwidth and bringing processing closer to the data itself. So that requires different types of memory architectures. And this is an area that we think is going to be very relevant in the future and where we invest our R and D dollars as well.
So I thought it was very important for this community to understand that above and beyond, the growth that we generate with our IP cores and buffer chip, we also invest in the future of memory and security because we think it's going to be really relevant for us. This chart shows the number of patent grants and the number of publications, which is patent request. We don't have the grant yet, that we've had over the last few years. And you can see the lower part of the stack bar is the patent grants. And you can see a small decline over the last 3 years.
But what you can see as well on the euro part of that, the top part of the side bar, is the number of grants, the number of publications, excuse me, the number of publications is increasing rapidly. As I said earlier, as we refocused our R and D dollars on 2 relevant technologies, we have increased our publications. To enrich our patent portfolio. It's very important for us. What I would add to this is that our current patent portfolio is still relevant because patents have a long lifetime, but it's also important for us to continue to feed, that portfolio with new grants.
We are we have about 202,600 patents. And or patents applications, which makes us 1 of the largest, you know, owner as an independent company in that space. The other thing that I think is important to understand because we had that question many times is so where do these patents apply? And why are they so relevant to the industry? And there's a confusion sometimes that we develop patents in the core memory technology And that's not the case.
We developed patents in the memory interface, in the interface technology. So if you look on the left side of that chart, here, sorry.
If you look at the
left side of that chart and these yellow boxes, they describe where we develop our patents. So this is the interface DRAM or the interface to flash. We have more than 1500 patents applications related to DRAM and Flash interfaces and the types of technologies that we address They span from DDR3, DDR4, DDR5, LPDDR3, LPDDR4 LPDDR5, GDR5, GDR6, HBM. So it's much broader than the products that we have in our portfolio. We try to look at all kinds of memory interfaces from a patent licensing standpoint.
And what's nice for us is that whatever memory the industry is going to deploy, that memory is going to have a memory interface, and that's where we play. The next one is chip to chip interface. We have 500 patents and applications related to the serial links, The product relevance is for PCIe, 10 gigabit Ethernet, USB and display port. Again, because it sits at the interface, it's relevant to a lot of standards. So it's relevant to a large number of companies in the industry.
And finally, in the security space, we have more than 100 patents and applications related to security and DPA. And the product relevance are in all of these standards that are needed, to provide a secure solutions to the market. I think it was very important that we clarify that because it's extremely important to our future. And looping this back to what we talked about earlier in terms of relevant market, when there's an explosion of data, there will be an explosion of interfaces to that data. This is a megatrend.
And our ability to continue to develop a patent portfolio that applies to all of these interfaces is extremely important for us when we see that megatrend The other thing that we are aware of is that when there's an explosion of data, there's an explosion of risk for that data to be compromised. So it's really, really important for us to continue to invest in foundational technology that protects data. And that's why you've seen what you've seen over the last year, us refocusing our company to memory interfaces and embedded security for the IP we provide, for the chips we provide, but also for the patent portfolio that we develop for our licensing program. This is a nice chart, but I wanted to show that chart because, it's a description of the types of technology we touch when we have a patent application. And the reason I wanted to touch to show that chart is that if you look on the left side, We touch every part of the memory system.
We touch the memory devices, the control signals, the data signals, clock signals, the memory controller, the IC, every part of a memory system is something that we have invention in. And I think this chart shows very clearly that although we only have only 2650 patents and that which is the large number, they're very, very focused in a very specific area. And that area is memory and security because we think that these 2 megatrends are going to drive growth for us. This chart is interesting as well, the art aspect of it is very interesting, but this, basically, this shows this is a representation of the citation that these companies have when they ask for a patent. They have to cite companies that they think you have prior art.
So for example, this is a number of times that Intel cited Ramas as a potential prior arts company in that space. And to give you an idea, the microns, for example, this is the size of about 400, I think, 400 to 500 citations. But what you can see here is that every company that is relevant in that space cites Rambus as having potential prior art in that space. So we are recognized as being leaders in those spaces. So in summary, we ramp bus, I keep saying this, we move and we secure data, and we think it's going to be extremely important given the megatrends that we see in the explosion of data.
Used by humans, but used by machines as well, and the need for embedded security. So we're focusing on our core strengths in those spaces. Our strong patent portfolio has continued relevance and we continue to invent and in those spaces. We have a predictable cash generation that we reinvest in R&D And M And A, as you know, last quarter, we generated $38,000,000 a little more than $38,000,000 of cash from our operations. And we're delivering, to high growth market segments.
So that summarizes who we are. And I'll be happy to take questions after Rahul talks to talk to you about our financials.
Luke, thank you again for the overview. What I'll do is I'll go through some slides that help provide a little more color about our financials We'll start with this slide, slide 34. So we are reaffirming the guidance that we provided for Q3 So you don't have to do too much of the mental math. This is exactly the same numbers that we had when we reported our Q2 results and guided for Q3. It's been a great quarter for us.
And I've been really excited about the traction that we've had in all aspects of our business. Obviously, there's still a couple of weeks left in our quarter, so we'll see where we end up in our range. One point to remind you is that the guidance that we provided for Q3 excludes the results for the payments and ticketing business, that we're in the process of divesting to Visa. So, I know Kit and JMP are working mostly on trying to close that as soon as we can. It's not clear whether we'll be able to close it by the end of Q3 or not, but working closely.
But what we've done is we've taken out any impact of the payments and ticketing business. Now as a reminder, for that business and many other businesses that are in the process of going through a strategic change, what we see is customers essentially pausing to do a new paper with Visa. So from an accounting perspective, you know, we'll still have the same rough expenses on a quarterly basis that we would expect, but I'd expect our revenue in that business would be lower than we would have done otherwise. When I report, results for Q3, what I'll do is I'll give you those numbers separately so you'll see what the revenue and expenses associate with payments and ticketing is so you can exclude it from our results. But again, we're reaffirming the guidance that we provided about Q3.
Let me talk a little bit about our revenue and licensing billings. For those of you that know us well are aware, the new accounting standard that we've had to implement last year called ASC 606 had a substantial impact in terms of how we report our financial results. It hasn't had highlights how strong that cash flow and predictable how predictable that cash flow stream is. So just as a reminder, in our previous accounting standard for ASC 605, if we were to sign a new customer contract that was $400,000,000, so $25,000,000 a quarter for 4 years. We would recognize revenue ratably over the course of those 4 years.
So every quarter, we would recognize $25,000,000 we build our partners and then collect it from them. ASC 606 changes that revenue recognition standard. What ASC 606 does has says that If you don't have a performance obligation in that contract, you take the entire value of the contract in the period in which you sign it. So that same $400,000,000 contract, the period in which we signed the agreement, we take roughly $400,000,000 of revenue We'd put the balance that we hadn't yet collected on our balance sheet as an unbilled contract, receivable. And then every quarter, as we bill and collect from our customers, see that receivable come off of our balance sheet.
Under ASC 606, what we would also do is record interest income every quarter. That's the imputed interest that we've given to our partners for those extended payment terms. So it's been a big change for us. And what I would tell you is that both internally as well as the analysts who cover our thought. What they've done is often just replaced what we had historically reported as royalty revenue, so revenue associated with those partners and contracts with what we now report in licensing billings.
So if you look at this slide here, Slide 35, same slide we published at the, earnings last year, what you'll see is that in 2018, as we were going through our transition, were able to publish numbers both under ASC 606 as well as 605, we were required to show that reconciliation between 606 and 605. And what you see is those numbers side by side. So royalty revenue for us under ASC 606 for 2018 was 130,000,000 However, what you see on ASC 605 is had we still been under ASC 605, it would have been a little over 300. Now if you go down to the 2nd series of bros there, what you see that our licensing billings for fiscal year 2018 was about 3.01. And that compares very favorably to what we reported as royalty revenue under ASC 605.
So again, from a cash flow perspective, our actual cash flow mimics what we had reported on licensing billings, and that's that we use operationally to gauge the performance of our company. What I wanted to do is actually spend now a little more time talking about, what Luke had described earlier in terms of changing the way that we think about our company and talking about our company as well. One of the things that attracted me to Grambus 3 years ago was these predictable long term licensing arrangements that we have, where we have no performance obligation and we'll continue to collect 100 of 1,000,000 of dollars on an annual basis because of our strong admin portfolio. What you see on the left chart is what we report as licensing billings, again, that operational metric. And what you see is that operational for billings staying roughly flat.
What I've talked about historically is we have the backbone of our business related to our IP business. Which is about $300,000,000. That's typically been about $250,000,000 from an architectural license perspective and about a 50,000,000 chunk from our silicon IP business. Now what I think you'll start to see in the coming years is that 300 continues to stay roughly flat will hopefully continue to grow as our Silicon IP business grows. But you'll see some structural step downs in licensing billings that we have as you see that Silicon IP business grow.
Now if you look to the right, what you see is very strong growth from a silicon IP and a product perspective. So when we call your products, it's primarily, what we sell as buffer chips. Now as you know, there is product revenue that we have, a small amount that we have that's related to our security business as well. So there's some workstation that we sell that are counted as part of our product business. But predominantly, what we report as products is what we sell from a chip perspective.
Then the next bar is what you see in contract and other. And again, the majority of the contract and other business is really Silicon IP. There is about $20,000,000 a year in Architectural Licensing or Licensing billings that really is more closely related to our silicon IP business but again, because we don't have any performance obligations, we don't recognize it as revenue under ASC 606. But again, contract and other bucket is really Silicon IP, which is our security course, as well as IP course. And then what I've shown also in the, the white box with is around it is businesses that are no longer part of our portfolio.
So for example, RLD was a lighting business that we shut down at the beginning of last year, in Q1 of twenty eighteen. And then our payments and ticketing business as well. As I mentioned previously, payment ticket business was about $25,000,000 to $30,000,000 last year. We expected it to grow to about $35,000,000 or $40,000,000 this year. But what you'll see is that as those businesses go away, the structural growth that we have in our product revenue will help us both from a top line and bottom line.
As Luke talked about earlier, what we see from a chip perspective is market share gains. As you see new platforms being introduced. And there's still a lot of, time left with with DDR4. We're consistently able to grow our market share in our chip business. And that's a very nice business for us.
That business has about a 50% or 60% gross margin for us. But what you find is that Now that our investment there is stable and predictable, as you see operating leverage. So while we ship additional chips, they may be a 50% or 60% gross margin but I expect them to be a 40% or better operating margin. Then what you also see as well is through continued engagements, both in security as well as memory, a growth in our cell IP business. And one of the things I talked about a few quarters ago is, as we did have a bit of a step down in our silicon IP business, specifically for our anti account bidding business.
We had a one time true up both in terms of actuals as well as forecast, and that's why there's a little bit of lumpiness here on a on a quarterly basis. What we expect to see is you expect to see both of those growing nicely out into the future and we'd like to continue that growth trajectory. Fargo, next slide, it really is about our balance sheet. We have a very strong, very clean, balance sheet. So we ended Q2 with almost $338,000,000 of cash, very small amount of debt for in our size.
We have 1, convertible debt instrument that's due in Q1 of 2023. We did have a call option on that convert. So it's only dilutive to us when we reached $23.30. So very strong, clean, balance sheet, from a contract asset perspective. As I mentioned earlier, with ASC 6 6, what we record on our balance sheet as an unbilled contract asset associated those licensing agreements for which we don't have any performance obligations.
So as of the end of Q2, we have almost $600,000,000 on our balance sheet. And again, that's discounted to present day value. We have a discount rate associated with each of those contracts for agreements for which we have no performance obligation. So one way to look at it is that we have over 3 $1,000,000 of cash and almost $600,000,000 of assets for which we have no performance obligations. And as Luke mentioned earlier, very strong cash from operations.
We've listed quarterly cash from operations for the last five quarters. Obviously, any given quarter, you'll have some variations terms of just what happens from a receivables or payments perspective. But what I said earlier is that I expect for us to generate over $100,000,000 in cash from operations this year. Go next, just a little more detail on cash. One of the benefits of our company is we have relatively low capital expenditures.
On an annual basis, it's typically $10,000,000 to $12,000,000, We'll have a little more at the back half of this year and early next year. As we've announced, we have a new headquarters facility that we're moving to a few miles away from our building. And so obviously anytime you move into a new facility, there's a few $1,000,000 of capital outlay associated with that. But one of the things that's very attractive about Rambus is that predictable revenue stream and that visibility in terms of cash flow. What we've also done as a company has been fairly consistent about returning capital to shareholders.
So through 2015 2018, we returned $200,000,000 of cash to shareholders through accelerated share repurchase programs. Our practice over the last several years to return 40% to 50% of our free cash flow back to our shareholders through agreed purchase. Obviously, right now, we found uses of cash from a strategic perspective allocation. The first thing that we look at is organic growth. So investing in product programs, which are growing very nicely, whether it's chips or silicon IP, also investing in R&D, making sure that we continue to invest in our patent portfolio as Luke showed, earlier.
Then next is inorganic growth. And you've seen that with the announcements we've made recently, with our acquisitions of both North US Logic, which we have closed, as well as their matrix, which we hope to close in the coming months. And then the 3rd priority would be returned to shareholders. To make it a little easier for you to gauge the ins and outs of the company, here what I provided is a little more context in terms of our So this is business. Again, we hope to close, eminently, kit and Jay are working on it.
But what we expected is in the second half of twenty nineteen, that business would have been $20,000,000 to $25,000,000, and we expected in 2020 it would have grown from 35 to 40 this year to 45 to 50 next year. So for those of you who model us, you would necessarily take that out. That payments and ticketing business was roughly breakeven for us, in terms of our internal projections. So you take a corresponding amount of expense out as well. Both the back half of 'nineteen as well as 2020.
Then we have in the middle is the Northwest Logic acquisition, which we have closed, I expect that to add about $10,000,000 of revenue for us in 2020 in this Q3 because it's close. We'll get the prorated amount expense. But as you know, through acquisition accounting, we can't count any contracts we assume as revenue, so you wouldn't get the same amount. But I do expect it to be roughly $10,000,000 of revenue for us next year. For the business that we are to acquire from Veramatrix.
Next year, I expect that business to be roughly $20,000,000 of revenue for us. It's very nice healthy business. Our focus, from an M and A perspective is, 1, to make sure that we continue to take care of our customers, also to make sure that we continue to take care of our employees and the partner and the ecosystems. So these are both very nice to our pro form a EPS in 2020, but it really is going to depend on just how the contracts, play out and the timing of the close for the business from Fair Matrix. What I wanted to do here on Slide 40 is to give you now a little more guidance terms of how we view 2020.
I know we've had a lot of parts moving in and out. So the previous slide showed some of those ins and outs specifically with payment ticketing going away and then us acquiring the business from Northwest Logic as well as from Vermaatrix. What you see on the left is licensing billings against our architectural licensing. So what we expect is that next year that business from a billings perspective, we should bill between $220,000,000 $240,000,000. As a reminder, about $20,000,000 of that I think is really related to our Silicon IP business.
It's just we don't have a performance obligation associated with it. The second is what we expect to recognize as revenue. Under our reporting, we call it contract and other. So that's really the Silicon IT business and expect that business to do between $70,000,000 $90,000,000 for us next year from a revenue perspective. I think there's absolutely opportunities for us in all of these businesses to do better.
As well. And then third is what we have reported as product revenue. Again, that's predominantly what we sell as buffer chips. And I expect that business to do between $75,000,000 $95,000,000 for us next year. So again, that's predominantly our buffer chip business.
Sometimes we do have some product sales in our security business as well. But I think our we see our chip business growing very nicely. With each new platform, you see a substantial improvement in terms of our design win momentum, and that's what gives us confidence that we'll continue to grow our revenue as well. On the right side of the page, we have some of our expenses and other pieces. So expect total operating costs, including COGS, be between $245,000,000 $265,000,000.
Now obviously, there isn't. Any COGS associated with our licensing billings, we do see COGS associated with our product revenue. And that's about a 50 or 60 percent gross margin business. Our silicon IP business tends to be about a 70% or 80% gross margin business, and that's really associated with the engineering teams that we have there. So total operating costs again between $245,000,000 $265,000,000.
I expect to see about $3,000,000 for us from interest income and expense because of the cash balance that we have. And I expect that we'll have on average over the course of next year about $115,000,000 of fully diluted shares outstanding. So again, this is our pro form a outlook. This includes royalty revenue of about $40,000,000 to $60,000,000. But I think most people who look at us look at the cash flow that's generated by our licensing billings income instead.
And what I'd like to do after that is to turn it back to Luke for some closing comments, and I invite the members of our exec team that are here up on the stage and we happy to entertain any questions that you may have.
Thanks, Raul. So as we said earlier, there are a few things that we've done over the last year. One was to simplify the company, refocus the into our core in semiconductor, high speed interfaces and embedded security. We consolidated that through set of acquisitions and divestitures. We divested our payment and ticketing business.
We acquired Northwest Logic And Verimetrix. We're also refocusing our internal R and D efforts onto the patent portfolio that we think is going to be relevant to our long term future. As we're doing this, we continue to generate cash. As I said, last quarter, we generated $38,700,000 of cash, if I remember well. We continue to do that.
And we are going to try to continue to simplify our business as we move forward and report our business in a way that would allow you to track us more easily. So this being said, I will leave the stage open to Q And A with