Ladies and gentlemen, the presentation will begin shortly. Please take your seats. Ladies and gentlemen, please welcome Oracle Senior Vice President, Ken Bond.
Hello, welcome to the 2018 Oracle Financial Analyst Meeting. Let's talk about what we're going to be showing today. Welcome. Mark will come up here shortly. We'll be speaking with you about our application ecosystem and the opportunities that we're seeing. After that, Mark will be inviting Safra up on stage. We'll have a Q&A, spend some time with you talking about the things that you would like to speak about. Somewhere approximately around 3:15, we'll be taking a break. Coming out of the break, Larry will be here. Larry will be speaking with you not only about cloud infrastructure and Autonomous Database, but basically the integrated suites of our application portfolio. Then we'll turn into a Q&A session with Larry. I expect the event will conclude somewhere around five o'clock this evening. You ask for it every year, I bring it to you again.
This is our most frequently requested slide by legal. Please indulge me. Statements in this presentation relating to Oracle's future plans, expectations, beliefs, intentions, and prospects are forward-looking statements and are subject to material risks and uncertainties. Many factors could affect our current expectations and our actual results and could cause actual results to differ materially from those. A detailed discussion of these factors and other risks that affect our business is contained in our filings with the SEC, including our most recent reports on Forms 10-K and 10-Q. Copies of these filings are available online from the SEC or by contacting our investor relations department. All information set forward in this presentation is current as of today, October 25, 2018. Lastly, Oracle undertakes no duty to update any statement in light of any new information. The second most popular slide.
We will be using non-GAAP financial measures, just as a reminder to everybody here. Lastly is just understand the presentations today are made for information purposes only. Thank you for your indulgence there. You won't be seeing these slides again. I just want to make sure everybody understood the things we'll be talking about. With that, I'll turn it over to Mark.
Imagine a world where everything is intelligent. A workplace where complexities are simplified. What if you could create tomorrow, today? Imagine augmenting human capabilities with powerful intelligence so systems know what you need, when you need it, on any device, personalized so you can take action quickly. Imagine using AI to source the best fit candidates. Instead of candidates finding jobs, imagine jobs finding candidates. They can access details via chatbots. New hires can complete onboarding activities in an easy and smart way, even before they start their new job, and be productive on day one. Achieve more than you thought humanly possible when human capabilities are augmented with AI. Oracle HCM Cloud, simply powerful.
Hi. It's a really upbeat audience. Good to have you all here. It's been a great week for us. We've had more than 60,000 people here. We've had an enormous number of customer engagements. I think you're gonna find I hope you had time to spend at OpenWorld going to some of the events, but this was probably as upbeat and exciting a conference as we've had in years. I think highlighted by what you're gonna hear from Larry a little bit later on the OCI and Gen 2 announcement. That was superb amongst a number of other really superb sessions. This, being with you, is the highlight of my week. What are my objective? My objective is to give you an idea about the applications business.
First, I'm gonna tell you a little bit about what it is and what we've done, just so we have a common baseline of the history of how we've evolved the business. Then I'll try to explain to you, given where we are, what the magnitude of the opportunity is. I'll try to lay that out for you in a couple of pieces. So this is just sort of, this is our revenue. FY 2014, a little less than $9 billion growing out to $11 billion. As we've gone through the transition, we went through a period of time where licensed revenue was declining as we were investing in SaaS growth. As many of you know, we did many things as we started generating SaaS growth. We did things like promotions.
We did things like really grading the volume of SaaS revenue that would come up over time as customers would go through their transitions from on-premise to the cloud. This began to change now as many of our customer base, we're now over 6,000 financial ERP customers in the cloud. So we've begun to generate scale, and it resulted in us, as you saw in FY 2018, generate more than double digits or a little higher than 10% growth. As we went into the year, I said I thought we'd have roughly double-digit growth for FY 2018. It turned out that we did, and I think we'll be on roughly the same track for that this year. So just a context of the size of the revenue. Let me just go back on one point.
This revenue, when we talk about ecosystems I'm trying to see if this clicker will move me back. It will not. Okay. When we talk about ecosystem revenue, it's got multiple components, and I just want to make sure we're clear so there's no confusion. It is our on-premise application support. It is our on-premise license revenue. This is different in tech, where a license is really portable between on-premise and the cloud. Apps is much more binary. You're either on-premise or you're in the cloud. This is our on-premise license, our on-premise support, and our SaaS revenue. All three components combined are now more than $11 billion in revenue. It's important just to make sure that you understand that's what's in it. This is what we've done from a portfolio perspective.
You can see the releases of Fusion, what we've tried to do is really highlight the different products that have come out, and certainly the major enhancements that have come out with each subsequent release of Fusion. By the way, just to make sure we don't confuse you'll see at the end of the chart, Fusion release 12, release 13, and then somehow it skips to R18. We'll now be moving to a convention where we'll talk about releases in the context of the years. It'll be R18 version one, which will mean year 18, quarter one. That'll be how the conventions of naming go forward. We've supplemented that organically built base with a series of acquisitions. You can see, other than NetSuite, the acquisitions have really slowed down, if you will, decelerated, and there's a reason for that.
We've got great strength now in the portfolio. We really don't have any holes in the portfolio now. Just to be clear, as ERP has evolved, manufacturing, procurement, supply chain, financials are all released. They're all on the market. They all have references. We have significant scale now in ERP and virtually every other category. As Larry mentioned in his keynote, we now have in HR, as an example, more customers than Workday. That's a customer count number to the best of our ability to count it. To give you an example, you see here this acquisition in 2012, Taleo. Taleo's got a large customer base. It focuses really on recruiting. As an example, in R13, we have now released HCM Recruiting. That's important because now that's rewritten as a Fusion module.
Now as we continue to mature the portfolio, some of those acquisitions are getting rewritten, those applications, so they're actually part of now an integrated suite. All of ERP is organic, all of HCM now organic. You see this, and this is particularly true in the back office, and you'll understand why I'm focusing on the back office in a couple of minutes. Okay. We think we have the most complete SaaS portfolio in the marketplace, and we don't really think it's close. In terms of having a back-office capability and a front office, we're really the only credible alternative in the front office to Salesforce. There's no other real alternative, and we are the leader in the back office. I'm going to focus a little bit back in history as to all of what we've done.
By the way, this won't come close to tell you all the things we did. I tried to capsulize a few points so you'd understand how many things we've touched in the business over the last several years. First, we realigned the sales force. You can imagine how popular it is every time you realign a sales force. We used to have a sales org that would show up and see the CIO and sell a bag of apps. The CIO, because of the architectures, in many cases, had maybe not complete control, but most of the control over the app decisions. As you move to cloud, that changed. Because the infrastructure and all the underpinnings of the infrastructure no longer become relevant in the decision process, because of course, it's cloud.
The buyer, the functional owner, now gets a bigger vote. We took our organization, our sales organization, aligned it by buyer, aligned it by product, and in many cases, by competitor. Our HR sales force calls on the CHRO. Sales, HCM. Competes with, most of the time, Workday. Our ERP sales organization, I'll talk to you in a second, is aligned primarily to call on the CFO and the organization around the CFO, selling our ERP portfolio, and competes with, depends who, I'm going to go through that in a fair amount of detail in a second. That's how the sales force is lined up, buyer, product, competitor. In terms of reach, we built out a global series of sales hubs.
Long-held belief that selling now, because of technology, can be done out of hubs, meaning that we can have an individual give six demos in a day over the network, as opposed to having to show up individually to give a demo one at a time, which was sort of our old model. Hubbing our capabilities became an important strategy, and we've spent the last several years building out that hub capability. As a result of the hubbing capability, we also altered our structure in a mix between inside sales and outside sales. It's a significant move for us because now, particularly with cloud, our ability to go after the market changed. Our ability now to sell to mid-market, smaller customers than historically we'd ever sold to, now increased. We would never have historically, called on Lyft. Lyft didn't have a CIO. They didn't have a data center.
We wouldn't have been calling on Airbnb. Both, of course, are now customers. Both are customers in the cloud. They frankly were born in the cloud, now with cloud, we get the opportunity to move significantly down market, so that this capability of inside sales and this hubbing has also led us directly into a new market, incremental adjacent market opportunity. Simultaneously, we've had to add things that we historically never knew about. Renewing a sales contract turns out is an interesting process. Turns out the length of a contract is an important dimension. How you renew, how you line up now because, of course, we deliver now instead of a product. Historically, we landed a DVD, a CD, gave it to a customer, they hand it to a systems integrator, loads of customizations occurred. Now we turn on a service.
That service has to run continually, 7 by 24. We have to upgrade that service, et cetera. The success of the customer, if you will, customer satisfaction, customer success teams, as we call them, become critical to the customer and become an important part of our ecosystem. Neither of these teams existed six, seven years ago. Not only do we have to change the sales force, build out hubs, change the mix of inside and outside sales, but we had to create incremental teams of people with capabilities we didn't have before to be able to pursue these opportunities. In terms of money, we obviously expanded the sales. This was really popular five or six years ago as I began to increase the size of the sales force. I'm laying on the sarcasm as thick as I can.
You looked at our OpEx, and I would get questions all the time. "Your OpEx seems to be going up significantly." Yes, it is. Why? Because we don't have enough capacity to serve the market. Now, what's happened as those of you that have followed us over the past couple of years is you've seen that OpEx begin to flatten out. Let me explain to you why, as it relates at least to applications. The application sales headcount in the company is up 60%. Again, it's a wavy line. I always tell Ken, if it's 62, don't call me and say I underestimated, or 59. I'm roughly right. The ERP HCM back office headcount has doubled. We have 2X the number of people that we had since just FY 2014. Okay. A huge investment for us.
The organization is supplemented, and there is a mix between inside and outside salespeople. The way we do that is we mix even with the headcount increase, as that mix adjusts, so does our OpEx. As we've moved resources down market, I want to make sure I'm clear, we are not abandoning the enterprise market. This is a supplement. This is an and, not an or. As we've added resources, the cost per seat of that resource comes down dramatically. The mix of those cost per seats, that field rep and that field SE compared to a hubbed salesperson and a hubbed shared SE, is dramatically different. Our market coverage has gone up dramatically since FY 2014. Yet when you look at our OpEx, our OpEx looks like it's up 1%, 2%, 3%.
That's because of the way we've mixed the resource over the past several years. We've gotten a hell of a lot more coverage, particularly since FY 2014, without having to spend a lot of money. All right. Next chart. Oh, I control the charts. A-ha. All right. Let's talk about sort of where we are now and where we're going. If your question about why customers are even moving. Do customers really want to move? Because I think years ago, there'd be thoughts that some of these systems would stay in place forever. Why would I ever change an HR application, why would I ever change an ERP application? Let me do my best to try to explain it to you. First, it's just innovation. There's just too much new stuff. First of all, you get four releases.
This says two to four, but in general, we are hitting about four releases a year per application. Let's stick with ERP. In ERP, you get four releases. Those releases could have a couple of hundred, 250 features per release. This is not like the old days of three or four years in between releases, where you have to have a systems integrator customize it. These come as part of the subscription price. I now get thousands of new features of functionalities that I didn't have before. Second, the user experience. We have gone way away from green screens and all of this stuff to now instinctive, intuitive UIs that can now be deployed across an organization. This is a brand-new user experience. The integration, you are going to hear Larry talk about.
The integration of machine learning and AI, chatbots, digital assistants, capability to get levels of efficiency and levels of innovation, the ability to crunch massive amounts of data and get them in front of an end user in milliseconds that I couldn't do in old applications. These are dramatically different. Now you are going to see us introduce analytics that are actually embedded directly into the app itself, as opposed to having to extract data, mart it, warehouse it, whatever term you want to. Those analytics, instead of the data going to the analytics, the analytics go to the data. They are embedded directly into the app. Customers love this stuff. Shows up in our numbers, shows up in all the references, this is an innovation. If there was nothing more here than innovation, customers would move.
The good news is you actually get to spend less while you get it. The costs simply go down. The cost of implementation is dramatically different. By the way, this is one thing you all will see in the marketplace. One of the things changing in the ecosystem today is the SI role. As opposed to staying with a customer for a decade and charging millions of dollars, this is now all about installing a pristine version of the software with as few customizations, or in our case, extensions. In some cases, you see the go-lives as quick as five, six months. Not three-year implementations, four- or five-month implementations. We automate part of that process, way driving down the labor cost.
Now when I get an upgrade, I no longer have to go repay the systems integrator again to feature string those customizations they built the first time, which by the way, was inherent in the old model. The cost of the upgrade is built into the subscription price. Generally speaking, we tell our customers to think of the overall TCO once the storage is gone, the data center's retired, server's gone. You got about 25%, 30% lower TCO. Imagine I could get all that innovation and spend less. By the way, if there was no innovation, I just told them you could save 25%, 30%, they are doing it. If I just had either one, I have both.
The fact is, we talk about security now, instead of that problem being the customers, where they have to patch, they have to do all of this complicated work, all of that gets moved to us. When you hear about Gen 2 and OCI, when Larry talks in a second, all of those benefits accrete to our application. Our entire applications business will run on autonomous database. Immediate patching, immediate optimization. All of that work will be done by us as opposed to the customer. Value proposition is strong. Market. You can get a view of the size of the market and the back office. The back office is roughly twice the size of the front office. Both big markets, $42 billion in the front office. This is not my numbers, by the way. This is IDC's numbers. They're very good.
As you see here through 2018, they are very good at predicting the past. Not quite as much when you look into the future, but if we look at the past, they're pretty good at the counting part. I feel these are roughly right numbers, and we would agree with their view of it. You can see the growth, with the back office growing $22 billion over the course of the past five years, $15 billion, $16 billion, $17 billion in the front office. This gives you a little bit of view of the applications that are contained in each group. Okay? That gives you roughly the sizing. That'll become important as I take you through a little bit of math in a second. Oracle back office. This is talking about us, this being the back-office applications. For us, let me define back office.
This includes Fusion ERP, NetSuite, and HCM. That's how we, Oracle, think of the back office. Currently, more than 50% of our application's ecosystem revenue is in this back-office part of the application. When I say 50, just to be clear, it's not 51, it's not 52, it's not 53. It's greater than 50%. The bulk of what's in our on-premise support is ERP and HCM. Okay? That's the bulk of our support. The back office, obviously being twice as big, and our presence being materially bigger, is key to our application's ecosystem growth. The opportunities fall in a couple categories. I'm going to take through this in a second. First, taking our install base, our existing customers, they could be E-Business Suite, they could be J.D. Edwards, they could be PeopleSoft, all back-office applications, and moving them to these modern applications in SaaS.
That has a significant revenue impact for us, and we think great things for our customers. The addition of new customers from, I'll just call them legacy on-premise vendors. We have another term here. I'll take you through who they are, and I also want to make sure you understand the size of the market that those vendors have, because I think it's going to be bigger than you might typically expect. You don't have to believe me. I blame this all on handlers, legal department, et cetera. I can't really blow up this chart according to Gartner the way I'd like to so you could read it. This circle is meant to just show you where we sit in their "Magic Quadrant" for cloud core financial management suites for midsize, large, and global enterprises.
You can see when you get out there, it's really not close. By the way, if I could, I think there is a This might work. Might this work? Yeah. It's not readable. I blame this all on Gartner. If you look at these names down here, you'll see there's SAP Business ByDesign, there's SAP S/4HANA, Sage Intacct. These are a whole bunch of companies down here. I just want to make sure. Again, one more time, this is not my opinion. This is Gartner's opinion of what they see on their vision and ability to execute. When you see this sort of gap in a Magic Quadrant, it's a big deal. Obviously, we look at something like, I don't know, Judy, 70, 80 Magic Quadrants probably across Gartner.
Very rarely do you see this sort of gap and this many people back over here. If you believe this market is real, this is a big deal. Let me just touch on NetSuite for a second, because obviously we spent $9.3 billion buying NetSuite a couple of years ago. I wanted to give you a view of what we've done with NetSuite and what the implications have been. Then I'm going to take you through how all this kind of comes together. Reach. We've taken NetSuite into new countries. No surprise, but NetSuite was very good in the United States and didn't have a lot of distribution outside the U.S. We focused very much on getting them outside the U.S. We've also spent a lot of time on R&D, localizing the product for Germany, France, China, Japan, Brazil, India, Mexico.
Material markets, where with a little bit of R&D, we've been able to get many of these things done in their last couple of releases. Localized product for more markets, increased distribution in those markets. Now, in addition, we've also increased their distribution in the United States. As good as they were in the United States, I'm sure you're going to find this shocking by my previous statements, but I thought they were under-distributed in the United States. The market is materially bigger than they were addressing. We've done two things, increased the amount of R&D, which has led to this ability to now get a localized product for more markets, simultaneously increase their sales force. We've added, for example, just to give you an idea, in the last 12 months, we've added 3,000 new customers. There's 3,000 new customers in 12 months.
We also now have the opportunity to upsell with NetSuite, planning, budgeting, SuiteCommerce, all of this. Some of this stuff is coming actually from Oracle, like planning. Some of it is coming from R&D inside NetSuite. Giving them yet a broader suite of products, more countries, more salespeople, more products. Simultaneously getting into more verticals. NetSuite has had great success going into not even verticals, but even micro verticals. Going after markets like campus bookstores, shoe stores. They've had tremendous success going after micro verticals, it's worked. They deliver an integrated sort of a template for implementation that dramatically lowers the cost of implementation. All right. That gives you an idea about NetSuite. If you look to the right, the NetSuite revenue has gone from 16% in Q3 to 22%, this is revenue growth, in Q4, to 26% in Q1.
This is a result of significant bookings. The bookings growth in Q4, our fiscal Q4, was over 70%. I want to make sure I said, it's not seven, I didn't say 17. Over 70%, driven by both the U.S. and international. Grew again in Q1 over 40%. These are numbers that, frankly, have been a result of everything that I've described. All right. Let me switch you to just a little bit of math. I got three or four charts, and we'll close. I'm okay on time? Okay. Opportunity 1. I want to take you through this slowly, not to the point that you get bored, but to the point that we get a lot of clarity. We have a $6.5 billion on-premise applications business today. That is a subset of the applications ecosystem I showed you earlier.
If I confuse anybody, please ask, because I want to make sure you don't get confused. We'll convert that to SaaS. We get three times the revenue when we convert that on-premise business to SaaS. We triple that revenue, and as we triple that revenue, the applications ecosystem in totality doubles. Anybody here who says, "That completely confused me. You went way too fast with the numbers." That subset that's in the on-premise triples, remembering part of what's not in that $6.5 is our existing SaaS business. When you then triple that and add the existing SaaS business, the ecosystem doubles. So the $11 billion becomes $22 billion. That simply is, the leap you have to take with that is that the E-Business Suite customers migrate to Oracle SaaS as opposed to one of those things in that lower left quadrant that Gartner has.
To believe this won't be true, you'd have to believe, "I'm staying on-premise the rest of my life," or, "I'm going to convert E-Business Suite to somebody in the lower left." I think very likely we will get all of these customers converted. Just to give you some context, we've moved, an inclining amount of our base. We're still just a little bit underneath 10% moved. When we say moved, that's moved core to core, meaning E-Business Suite financials to Fusion SaaS financials. We've got now 35% of our base that has bought something in the cloud from us of that base that doesn't replace that core. We have two-thirds of the base in some stage of our pipeline. Our base is engaged. Hasn't all shown up in revenue here, but our base is highly engaged. Let me take you through category 2.
The first thing I want to say is the people that I generally talk to think Oracle and SAP dominate the back-office market. The truth is, when you add the two of them up, it's less than 50% of the market. It's 48% of the market. You come back and say, "Well, who the heck has the rest of it?" It's a hard question to answer because it's just a blizzard of companies. Everybody from Deltek, Tyler, Unit4, Epicor, Sage, the blizzard of companies that are under the There really is no Infor. There's a Lawson, there's a Baan we didn't even get them on the chart. There's an Infinium. I could go on and on. This is a blizzard of companies that have 1%, 2% market share. Many of them are private equity. Nothing against private equity.
They may be through their second or third pass through private equity. They've endured the, "Let's spend less money in R&D. Let's get cash flow up." I always think that's a great strategy for all of them. I think they should lower their R&D, lower their sales expense. These are mostly the companies that are in that lower left of what you saw on the Gartner quadrant, if they've even able to go get a cloud product. What you would find is that when many of these companies say they have a cloud product, what they got is like one module that's in the cloud, or they've taken their existing on-premise application and hosted it at AWS and call it cloud. As opposed to rewriting the app for pure SaaS with all the modern capabilities that I've described that are in our applications. That's who's in the base.
I am not predicting this as an outcome because it's a pretty low outcome. I wanted to give you an idea of the magnitude of the opportunity. All I've done is taken the $35 billion back office, and what I've done is taken SAP out. I took them out. I took Oracle out, gone. The remaining part of this is $35 billion in the hands of that blizzard of companies I described earlier. If only 20%, and only 20% happen. By the way, I'm going to stop. If only 20% move at the same multiplier, which is roughly what we've seen. By the way, we're the beneficiary. As I've told you, 50% of what's in our SaaS revenue today has come from some form of our base.
Even though it's a small conversion, it's come from some form of our base or our customers, in terms of number of customers. Half of our SaaS revenue has come from outside our base. We have deep experience in migrating these customers. With 20%, if you look to the right here, the opportunity for us would be an incremental $21 billion. You can decide whether you think, "Well, he's wrong, it's 19, 18.5, 20." Invent whatever algorithm you want to. This is roughly right. The answer isn't $12 billion, and the answer isn't $30 billion. If you get 20%, you're going to be roughly close to the number that I described. By the way, that comes from lots of conversions we've done, reasonable extrapolations we can make from the existing work that we've already done.
The reason-- Well, I'll talk to you about that more in a second, but you can see how big the numbers if you get. For us to wind up with 50% share More than-- which is what most markets eventually mature to, 20% plus moving our base would get us into the mid-30s eventually of overall market share. I believe we're so far ahead now, we'll do better than that. Let me see. I think I'm Am I missing a chart? Don't I add them together? I should add them together. I'm going to do it for them. This is advanced work that I've got to do. Okay. Yeah, I think it's in there. Let's go back here to the bottom. Thank you, Safra.
If you look at the bottom, the applications ecosystem grows, I'm going to stick on that, from $11 billion to $43 billion, there we've done the math, if you believe we're going to convert our install base. You can say, "I don't believe you. I don't think they'll go. I think they're going to go from the upper right to the lower left." Seems like a smart move. I don't think that's what's going to happen. If you believe we're going to get 20% of that remaining install base, I actually think we'll do better. The reason I didn't add it in was because the number got so high you would think I'm crazy. Well, you may think that anyway, but I would add to that belief. This does not consider the addition of any current SAP customers. Let me spend a second on this.
SAP is currently telling their customers they will end of life their current product by 2025. This will force customers to convert. We have customers now calling us that have never talked to us before on these applications, saying, "I'm faced with an $800 million bill to move to S/4HANA, and I don't know that I get incremental functionality as a result. I get a new UI. That's about it." I had a customer last week, and I wish I could have asked her if I could use her. She's faced with a name company, a $350 million bill. Her point to me was, "You have to come in here and give me. I have to go to my board and explain that I went through the work to look at my alternatives.
I need you to put a team on this, and I need to put a team on this because I have to do the due diligence." This decision they've made will put their install base at play for no other reason than the financial due diligence of the magnitude of spending they're putting their customers through. I have factored none of that into this, none. I've also not considered growth in SaaS sort of outside the back office. I think you're going to see over time, we'll show you more of what's going on in CX as it relates to our opportunities to continue to gain share. We've had a very good run in sales auto, as an example. Mainly because of, in fairness, some cases, the prices that Salesforce charges.
It's really actually not a bad thing when you're starting and you're a very credible number two, and you have a suite. By the way, as we increase our ERP base, the opportunity to sell CX products, customer experience products, into that ERP base inclines as well. I have also not factored into this selling incremental modules. For example, I mentioned Federal Express on our call, I think at the end of Q1. Federal Express had an EBS, E-Business Suite application that ran FedEx. They bought a company called TNT in Europe. They ran SAP. They're converting both to Fusion ERP. As a result of that, we got incremental modules that we didn't have under E-Business Suite as we expanded the entire base. I did not factor that into this either. ERP, again, I'll go back up to the top. The back office is the key.
It's twice as big as the front office. To be very blunt with you, as Gartner says, it's where we're the best. Okay. With that, maybe we'll go to Q&A. Safra? Safra's going to come up. We'll take whatever questions you like.
We'll be starting with Brad, I'll go to the mics at the back. If you have a question, just kind of hold your hand up and I'll get around to you. As will Eva.
Son, thank you very much, Mark. Brad Zelnick with Credit Suisse.
Hi, Brad.
Compelling presentation. I just had a couple of questions on the opportunity. If we think about the $11 billion incremental in transitioning the base over to SaaS, you've been in the business of making predictions in front of us before. Can you talk a little bit about what the rate and pace of that has been, what you expect it will be? If we look back three, five years from now, how much will have come over, and to what extent can you use carrot and stick to move them along? I've got a follow-up for you as well.
Yeah, multiple questions, which you're good at, right? I would say, first, we have chosen not to use the stick. I think you can understand the difficulty when you put a stick out there like SAP has, to basically say, "You're going to move or else. We're going to end of life you." In our opinion, this is not a great strategy. In fact, to be very blunt with you, our Oracle E-Business Suite customers are relatively happy. We've given them new releases. We just put E-Business Suite 12 out there, so they got some more features. In fact, we make a point of telling them we're not going to end of life their product. They're in safe hands with us. Now, what those customers are doing, Brad, is they're buying modules around the core.
I think it's very likely that what you'll see is a migration where if I'm an EBS customer, I may buy financial planning in the cloud. I may buy procurement in the cloud. I'll take my time as I go through this migration, and to be blunt with you, Brad, that's fine with us. One of the big things we're doing, I didn't mention this in the install base, is we're dramatically lowering their cost of conversion. We've released a program called SOAR, S-O-A-R. SOAR is basically a program where we guarantee the outcome of basically telling the customer whatever the subscription is, we'll charge you twice that for implementation. We'll actually accelerate your go live to five, six months, and we'll make sure you're done right. If not, it's on us. Our customer base, again, loves this. It really is carrot that we're driving.
In the end, Brad, just when I look at the pipeline, we're going to move the entire install base. If you're asking me for a quarter when it's all going to be moved, you can try that from Safra, but not from me.
I won't do that.
Okay.
Okay.
I would like to say that I think the big difference between this year and last year is that more and more of our customers are in fact moving their core. More larger companies, if you walked around, you'd see them get up and say they're live. You'd be hearing them talking about moving off the E-Business Suite or moving off PeopleSoft Financials, a lot of larger customers. That's the excitement when I hosted and Jeff Henley hosted Modern Finance Experience. That is a room full of people like me, who are responsible for their finance operations. I was sitting at a table that had small and big customers, some of them, like Walmart, had only moved planning and budgeting. They're a SAP customer, but that had gone unbelievably well, while other customers were moving the core.
FedEx got up and said, "We're moving the core." That's the kind of thing that is a difference. I also hosted public sector. Public sector states, full financials, FI$Cal from California, which is the finance back end, sitting on the other side of me, Texas, moved to Fusion. Across North Carolina, looking at seriously talking to all of them about moving the state's financials to Fusion in the cloud. That's the big difference we've got going this year. It is true many will still take transportation management and other parts, for folks like us, Oracle, Fusion Financials, all I need to do is invite Maria up here to tell you what it's meant to us and the capability that us, we're a pretty big company. We're global. We're live on supply chain and of course, financials. This is a big darn deal.
It's for real, it's making an enormous impact on the way we run the business, our goal is, like some of the folks at Modern Finance Experience, I don't know if any of you were there. You know what their goals are? One day close. That's their goal. That is now, if you ask Maria, she's going to tell you that's our goal, too. I know you're all trying to figure out which one Maria is. Maria, raise your hand, please. There she is. Okay. She reports directly to Corie West, our Chief Accounting Officer. She is the user for Oracle Financials. That's the big difference between last year and this year.
My quick follow-up. Thanks.
Before you follow up, I have one other thing. That's our base. These other companies we're talking about that have a lot of the legacy base, they are much more threatened than our E-Business Suite base. If you actually asked a rep, a sales rep, who's in the ERP sales organization, we organize our territories by competitor, so they actually have a set of these either traditional legacy vendors, or they may have an E-Business Suite territory. They would rather have the competitor territory because the competitor is highly threatened that there's no future, or the customer of the competitor, more threatened because there isn't a future. There is no cloud strategy. There is no real cloud strategy. Therefore, there's even a faster impetus for some of those to convert.
Safra, your shareholders, I think last quarter, were very pleased to see the company take advantage of the opportunity in the stock and to see you buy back $10 billion worth of shares. How should we think about the limits to which you will use the $60 billion in gross cash and your $15 billion some odd dollars in free cash that you generate to buy back more of the company? Thank you.
As I said on the call, our stock is an absolute bargain. This is from our point of view, in comparison. We always balance acquisitions, other uses of cash, but we have a lot available to us, and we know where our business is going. It is very clear from where we're sitting at the opportunity that is ahead of us. It is very clear to us that in OCI, which Larry's going to talk about, Autonomous Database, we are on the field now with a product that, frankly, has no competitor to run our Oracle production workloads. We have changed the game. This is going to get very, very scary for everybody who is on a Gen 1 Cloud. Running the Oracle Database in OCI in second gen is a totally different ballgame. It takes us a while for folks to understand. We've been highlighting it.
We've been introducing it. We've been sharing it sometime early on. As customers see the opportunity, it is unmistakable. Verizon at Leader Circle sat with me, Mahmoud. They use AWS, they use Azure. Oracle workloads, OCI. These are folks who have some of the most important and difficult workloads in the world. These are folks who are absolutely security paranoid. These are guys who need high performance, who need absolute enormous capacity, they understand the value. Not only do they understand it, they're willing to sit up in front of Leader Circle, in front of all our big customers, talk about it. You're asking me, what am I going to do?
I think our stock is an absolute bargain, we're going to break out, I don't know when we're going to break out, I'm going to balance our other cash needs, of course, I require approval from the board for different amounts of purchasing. As long as it's a bargain like this, we may never get this bargain again. At one point, people are going to realize what we have been hiding, sort of under a bushel here. At that point, it's going to be a lot more expensive. Until then, I'm all in, our team thinks this is a bargain of a lifetime.
Hey, over here. Ryan Walens from Practice. Hey, Safra. You didn't mention PaaS as part of the application story, if I go out and talk to the guys out there, they tell me PaaS is the hidden gem because you can basically use it as an extension, et cetera. Can you talk a little bit about the opportunity? How do you see PaaS? Is that just something that helps you don't really make that much money off it?
No.
No, nothing to do with that. I just didn't want to include it as yet another reason to goose up the revenues, right? I think that in PaaS, to give you example, in FedEx that I referenced and Safra referenced, the PaaS part of that transaction, which had nothing to do with the app, was 50%, 60% of the upfront transaction to basically do some extensions to the application. Of course, you're right. Our margins on PaaS are great. It's just yet again, like I told you, there are many things I didn't include in that analysis. I didn't include incremental modules to which, as I said, FedEx bought incremental modules. FedEx, again, using them as an example, bought an extreme amount of PaaS, as part of the initial transaction, I think that'll do nothing but grow as we go forward. Of course, you're right.
I think of PaaS and license and OCI as one thing. It is why we changed the reporting, because you bring your database licenses to our OCI cloud, and if you have got enough licenses, if you have got the right licenses, you get Oracle Autonomous Database, you get a self-patching, upgradable automatically database. When I talk about OCI, I am also thinking about those Oracle workloads which display themselves as PaaS, but for which the money may be coming in from using your licenses that you buy or your support, and that plus IaaS is PaaS. It is all the same. It is why we changed our reporting, because there is an arbitrage in there for which we are very happy to have our customers decide whether they want to rent licenses, rental PaaS, or they want to buy licenses and pay support and BYOL, bring your own license to OCI, still get PaaS.
Hi, it is John DiFucci from Jefferies. It sounds like what Safra is saying is this is going to be the year where we actually may see an inflection point in the adoption of financials in the cloud. We saw that in HCM a while ago. I am just curious, Mark, can you sort of gauge about how much is left? If you think about the enterprise market, which is sort of a zero-sum game, how much is left in HCM in the cloud? Also, if you can hit on something that is still evolving, and that is NetSuite, it just seems like Oracle, Safra is talking about these really important workloads, these big workloads. That is Oracle's history. That is what Oracle is. NetSuite is something a little bit different. Are there things that are happening behind the scenes?
It seems like NetSuite is sort of running off over here, are there things happening behind the scenes to sort of really go after that mid to smaller market with NetSuite? I know it is two questions. Thanks.
Yes, two questions with a bunch of different branches, I think, which again, I look forward so much to coming here. I would say again, NetSuite, and I probably did not do a great job laying this out again. When they were a public company, they were limited in many of the things they could do in terms of investment, rate of investment. To give you an idea of the magnitude I feel about NetSuite now that they are part of Oracle, in the U.S. mid-market, I believe they could quadruple. That is how big I think the opportunity is. They were limited. Let me tell you some of the things we did in NetSuite that were amazing. In fact, I am amazed because I thought it would be a great idea. I am really thrilled to see, obviously, the outcome that has occurred.
We took 100 of our class ofs. They were short on people. They didn't have a enablement program near as mature as ours. We took their content and overlaid it in our enablement program, put our Oracle hired class ofs, and in nine months, NetSuite got them productive in the mid-market. We don't get that done at Oracle. It takes us two and a half years to get to productivity. They did it in nine months. We're trying to give them, John, the best of Oracle and make sure they don't get slowed down by any other thing at Oracle that might slow them down. We've tried to give them our products to add to their suite, leverage our distribution in countries where they don't have an employee. They don't have an employee in some of these countries where we do.
What our international organization has done, has been able to accelerate their hiring at a level it would have taken them two years if I'd have told them, "Go get started on your own." We've given them an enablement program. We've given them more distribution. We've given them more R&D. Just as important, they've now delivered and released localized products for many geographies they didn't have product for before. When I say we increased our R&D, these are small numbers, John, relative to Oracle, but big numbers compared to NetSuite. With a simple change of $10 million, $15 million, all of a sudden, and you're in five new countries. These were opportunities, John. We've sold nothing. We, NetSuite, have sold nothing. I think the opportunity is gargantuan for NetSuite. You had another question about HCM.
This whole thing, HCM and financial, we're in inning one and a half. Remember, the way I think of HCM may be a little different than how you do. Let me try to give you some context. Where I think Workday has done well is in particularly large U.S. big companies that have a strong CHRO that typically doesn't have alignment to the CFO. As a result, when that becomes an isolated decision, where it's just an HCM decision, we bat about 50/50. They got a good product, and they've had a little different situation. I understand their strategy to get to financials. It just hasn't worked very well, as a product, which turns to be a limiter when you try to get a debit and a credit and all of that to come out and be able to close your books.
As a result, also, as you move down market, people start to look at these decisions as a common back-office decision. It's not two decisions, it's one decision. As we move down below these biggest accounts, many of our HCM wins now are coupled with an ERP win. We win ERP, you win HCM. I believe the market for that is just in its infancy now. We're talking about inning one, maybe top of the second, as it results to how much opportunity is still left.
Very good.
Keith Weiss from Morgan Stanley. Thank you guys for hosting us. Safra, a question for you. One of the hallmarks and one of the really strong points of the story over the past couple of years has been the expense discipline. You've kept OpEx growth really low. Mark's talking to us about a lot of investment, and it sounds like there's a huge opportunity that he wants to continue investing behind. Are you going to be able to keep that low OpEx on a going-forward basis and keep the operating margins going? Is there enough puts and takes within the P&L that you could keep funding his investment in sales and marketing without taking OpEx really much higher?
Yes. I'd say quite easily.
First of all of that stuff we've done, and Keith, by the way, I think it's a great question. Last two or three years, we've self-financed, to your point. Underneath it, we have changed the story significantly in terms of the mix, the opening up of the Safra and I have opened up hubs all over the place, but we've self-financed this as we've gone, and I think that you will see that continuing.
You see, we have the benefit of also having incredible margin improvement in parts of our business. There are things, just like you said, puts and takes. The SaaS business is now at scale. Throws off a lot of money for us to continue to invest in. Our capital investments also been very measured because they've given us back a lot simultaneously. We've moved around in different investments in hardware and things like that. We are getting economies of scale in some areas and focusing in and investing in others. You can see it in our operating margin improvement.
Just one example to Safra's point. If you look at as our SaaS business approaches $10 billion, you start looking at our margins in the 70% range today. I believe we have de minimis investments in our SaaS estates, much of those for technical reasons, things we can do now with technology that we couldn't do before. As our margins approach not just 80%, but past 80%, 80%, 85%, all of a sudden that starts kicking off $1 billion, $1.5 billion worth of incremental margin. These are big opportunities that we have. As Larry talks to you about OCI, we've got other things in the organization that will actually collapse around OCI to the good as we start to consolidate many of our platforms.
I think the operating leverage in Oracle is still significant, and we will fund this investment in roughly in line with what you've seen as the OpEx changes that we've had over the past couple of years, Keith.
Okay. Thank you.
You know what, Mark? We've got Larry here already.
Wow.
Maybe-
You know what's very impressive, on time
on time early.
Actually, I think you're a little early, which is.
Yes.
Yeah, it's like five minutes early. I'm not used.
We're a little dumbfounded.
Yeah.
We're thinking.
We're rolling.
The question is, do you guys need a break, or are you ready for Larry?
They're ready.
Ready for Larry.
All right.
If you got more later, we'll be back. Why should he wait?
I'm the last person on, is that right?
Yes.
All right. Okay, I'm going to go through. Someone cue my slides. I'm going to do my two presentations rapidly. I'm going to go through and highlight a few things that I told our customers and then try to answer your questions. All right. Are my slides any place? There they are. Okay. We decided we're going to go straight to Gen 2 Cloud. Everyone's kind of sitting around gen-1. You could say we're in infrastructure. I would say in SaaS, we started before anybody, in terms of SaaS. NetSuite was the first cloud company in the world, I believe. It was started by me and Evan Goldberg a while ago. It was the very first cloud company.
We built NetSuite, about a year later, Marc Benioff, who was friends with Evan, looked at it and said, "Hey, this is a really cool idea, doing accounting on the internet. We'll do Salesforce automation on the internet." That was really the beginning of the cloud business. Oracle started converting its applications from on-premise to the cloud with a project called Fusion that we started well over a decade ago. We've been doing this for a very, very long time, this Fusion stuff. The applications or SaaS is a much older business than compute and infrastructure in the cloud. That was started by Amazon, not by us. It took us a while to respond to what Amazon was doing, and we did things Amazon-ish, as did Azure, as did Google.
We had a variety of issues with the first generation of our cloud, what we saw as the first generation of how Amazon does it, how Google does it, how Microsoft does it. We decided just basically to start over, which we did. We started over, and now we've delivered this thing, now it's available now, called Generation 2 of the Cloud. What was the big deal? Why did we decide to start over? There's a huge problem with existing cloud architectures in terms of security and reliability as far as our customers are concerned. If you're building Angry Birds, doesn't matter. If you're piping in Netflix to people's homes, doesn't matter. Really, for a lot of applications, what we worry about doesn't matter.
When we worry that it has to be available all the time, can never go down, and you can never let other people steal your data, you end up with a very different solution than what Amazon is doing now and Microsoft is doing and Google is doing. For our customers, we had to build a completely different approach to infrastructure than they took. We were copying in Gen 1 of our infrastructure, then we decided, no, we had to do it differently. It's not just machine learning that we exploit. We are absolutely devoted to machine learning, by the way, in our infrastructure and our applications. In our applications, we use machine learning to automate complicated tasks like closing the books, fraud detection, lots of things. We do a lot of automation in our applications.
We use machine learning to provide a new voice UI to our applications. The way you submit an expense report with our applications now is you take a picture of your hotel bill, you're done. We figure out what the hotel bill is, where you were, what is associated with it. You just take pictures of your restaurant bill, your hotel bill, you're finished. We use a lot of AI in the application to automate tedious tasks that people do all the time, like fill out expense reports, all the way to complicated tasks done by experts called closing the books. We use AI machine learning in our applications. We also use AI in our infrastructure. We built the first Autonomous Database. It tunes itself. If it needs more compute, if it needs more network capacity, it goes out and gets it. It backs itself up. It recovers.
It detects threats and patches them while it's still running. Amazon doesn't have to do that. Their customers don't expect a bug to be patched while the system is still running. Our customers do. We just had to take a very different approach. We use a lot of machine learning, but it was more than that in terms of re-architecting our cloud for reliability and security. Next slide. Smartest people in the world are being ripped off every day. People at Google are smart. People at Amazon are smart. People at Facebook are smart. People in the government agencies are very security conscious. They're getting data stolen, masses of data stolen all the time. They're down a lot. A lot of these systems are down. A lot of Amazon cloud customers are down a lot. This doesn't work for us.
This is why reliability, security, these are the reasons why we had to re-architect our cloud. What we had to do was Let's look at the next slide. What we had to do was basically fundamentally change the hardware as well as the software in our Gen 2 Cloud. We added a whole new network, a whole new network of computers to form an impenetrable barrier about our cloud to prevent people from the outside getting in. I'll show you another picture that will kind of explain all of this. We had to isolate our cloud from outside threats. The bigger problem turns out to be, someone takes a credit card, rents a computer, we let them in. We give them a computer. They can load all the software they want into our cloud. They gave us their credit card, right? Come one, come all.
Some guy in the Ukraine sends us his credit card, which he stole the night before. We let him into the cloud, they're loading stuff, loading all sorts of software into the cloud. They're looking around. This is really kind of a tricky business. We not only have to erect a barrier, an impregnable barrier, an impenetrable barrier around the perimeter of our cloud. We have to create barriers around each customer, so those customers are isolated from other customers. You have to trust everybody that you're sharing the computer with. Who vets them? Who makes sure, oh, this person is trustworthy, this person's trustworthy. This person I really don't know. I got his credit card number. That's how I validate him. Give me your credit card number. Welcome aboard.
We had to create barriers around not only the perimeter of our cloud, we had to create a barrier around all of our customers, individual customers. We basically had to create an isolation zone if this is what they wanted. Not everyone cares about this. Not all of our customers care about this. Some of our customers who are writing code, who are developing new applications, they're not worried about getting hacked. They don't care if they're hacked. They want to get on as quickly as possible, want to be productive, want the lowest price. If it goes down occasionally, it's not the end of the world, right? Not the end of the world. If you're running an e-commerce site and it's the week before Christmas, can't go down, has to be elastic. You can't have anyone hacking in and stealing. You can't have other people.
You care about reliability. You care about security. You don't want to trust the other people. We had to do that. We had to create these barriers. That was kind of one thing we recognized we had to do. We had to isolate our customers one from another, and there was no way in the current architecture we could do that. Not really. I'll show you why. The next thing we had to do is never go down, never fail. When a threat was detected, we had to fix it immediately. We're being attacked. We're being attacked by botnets. We're being attacked by networks of computers that are doing denial-of-service attacks or trying to stick malware into our cloud, trying to do all of these things. Remember, you have a credit card, you come into our cloud. You make it through the perimeter.
We have to have our own fleet of robots looking around for these threats all the time. When we discover that it can't be human beings, it's got to be our own fleet of robots looking around for these threats all the time. When you find a threat, you have to be able to kill it. Now, killing it, let me explain what I mean by that. We find a vulnerability in our database, and we have to get rid of that vulnerability. That means we have to patch the code of the database. We have to change the database code. Someone's figured out a way to exploit the database or exploit our operating system, the Linux operating system. We have our own version of Linux operating system. How do you do that? Do you just say, "I'm sorry, we found this vulnerability.
We'll just shut down your e-commerce site, patch all of our software, and then we'll bring you up. It'll only take about 10 minutes or something like that. Don't worry." They say, "You're not taking me down." This is what happens. This is why this is a huge problem, because when people discover vulnerabilities on premise, they discover vulnerabilities and they want to patch them, there are a bunch of human beings that get involved, and they say, "Oh, I got to schedule a downtime window. I can do it. I can't do it now. I got to wait after the holidays. I got to find a spot where I can turn this stuff off, and I can patch it." Doesn't work. You can't have human beings. You're being attacked by robots. You can't have human beings in the defense chain.
You can't say, look for downtime windows as an opportunity to patch everything. The disaster with Apache Struts, where the CEO lost his job, was this vulnerability was known for months, and they did patch 80% of their Struts databases. They couldn't even find all the databases to patch. Human beings can't do this job. You've got to have autonomous robots. You've got to use a lot of AI. You're being attacked by robots. You've got to have robot defenses. You have to find the threats, and then you have to be able to kill them immediately without taking the system down. It's really hard problems. We built Gen 2. The next slide. I got two slides in front of me now, which is interesting. You got two slides also? Let's go to the next slide. The Gen 1, Gen 2. Here's really a picture of the problem.
Here's the Gen 1 Cloud. This is how it works. You go into Amazon, here's your credit card. You go into a shared computer. There are two problems with the sharing. One is there are multiple users or multiple customers in the same computer that you have to trust. That is not the worst problem. The worst problem is Amazon's code is also in that computer, and they have to trust you. They have to trust their users not to change their cloud control code. It's an impossible problem to solve. These are Intel computers. They don't have memory protection, hardware memory protection. I can come in, rent an Amazon computer. I can look around in that computer. I can find the Amazon cloud control code, part of the operating system, and I can alter it.
I can then go from one computer to the next computer, looking around, spreading out laterally, collecting data. I can do all of that. That's a huge problem. Our solution to the problem is we have two kinds of computers in our cloud. We have computers that we rent to customers, and the customer can choose to have one all to themself and not share with others. If you're really security conscious and say, "Okay, I'm going to have only my code. Only Bank of America's code is going to go in this computer." Are you going to be someone, look, I'm a programmer, I'm a startup, I'm writing code. I don't care. I'm happy to share the computer because I want the lowest possible price, and shared compute is cheaper than dedicated or what we call bare metal compute. Great. We offer both. You can share.
You can not share. It's slightly more expensive not to share. You're not going to share with us. There's no way we're going to put our cloud control code in the same computer that anyone with a credit card can come and just jump in to our cloud. We're not going to do that. This is why as we better understood all of this stuff and all of the problems, we decided to rebuild our infrastructure cloud with a separate set of computers whose responsibility was to do cloud control. Sometimes you call it the control plane. These are cloud control computers. Only our code goes into those computers. No user, no customer. We don't trust anybody to put their code into the computers that we use to control the cloud. No way. Now it's expensive. We had to add computers. We don't use Intel CPUs.
It's not a computer that's easy to hack because you can't even get at it. There's no real internet access to this computer. Stuff can flow. We flow messages through the computer. We encapsulate messages, flow them through the computer because we're sending messages to one customer zone or messages to another customer zone, and we're routing them through our cloud control computers. Those are just messages. There's no way you can get your code into that computer. That's the way we think we can create these impenetrable barriers, not only around the perimeter of our cloud, but also between customers who care about stuff like that. We had to do that. Everyone's going to have to do that. We decided that, okay, we're going to be first. This is our customer base. These are the people who buy the Oracle Database. They care about security.
They care about reliability. It's very important. They're banks, they're phone companies. They are people that are just reluctant to move to the cloud. This is why. That architecture picture is why a lot of people are reluctant to move to the cloud. There are issues. We decided, okay, we're going to be the first out with this new secure cloud system, which has these Star Wars defenses, impenetrable barriers, managed by a separate network of cloud control computers. All these autonomous robots looking for threats, killing threats without taking the system down. This is the bulk of my presentation. Next slide. This is just a picture of these. They're physically separate computers. They're cloud control computers, two networks of computers. Threats can't enter from across the perimeter. Threats can't spread from one customer to another unless the customer chooses to share.
The second you choose to share a computer, you have said, "I trust the other people in this computer," or, "I don't care." Next slide. Just something about nomenclature. Our second-generation cloud is not just infrastructure. We have this thing called OCI Infrastructure, Oracle Cloud Infrastructure, which includes things like the database and compute and storage and networking and all that jazz. In our generation two cloud, it also includes our layer of SaaS applications. All of our SaaS applications, if they're not there now, everything is moving to this Gen 2 Cloud. We want all of our SaaS applications to have the same attributes. Everything's backed up automatically. Everything sticks to automatically, never comes down. You can't steal data, all of those things, everything. We're building one kind of data center, these Gen 2 data centers that run our SaaS applications, run our infrastructure.
Our SaaS applications are enormously advantaged by running on this infrastructure. They're faster, they're more reliable, they're more available, they're more secure. There's a rich set of tools you can use to extend the SaaS applications. You have all the infrastructure tools. We're the only enterprise SaaS company that is offering competitive infrastructure. If compared to Salesforce.com, they've got this thing called Force.com. We've got our Gen 2 infrastructure. We're trying to compete with, beat Amazon, Google, Azure in infrastructure. I don't think you even think about S alesforce.com in that marketplace. If you're extending a Salesforce app, if you're adding a data warehouse to your sales automation system, how do you do it? You do it with Force.com. How do you add a data warehouse to our SaaS applications? You use the Autonomous Database and all of our cloud analytics, et cetera.
We have an enormously powerful infrastructure and tool set that can be used to extend our SaaS applications. None of the other enterprise SaaS companies have anything like this. They're not even in this business. You think of it as like, well, this is a totally separate market. You're selling sales automation or something like that. You're not in the infrastructure business. You're not in the database business. Really? We think these businesses are very closely related. It's true that Amazon did all of this stuff, and again, everyone says, "Hey, we're competing with Azure," and blah. We have one set of competitors for infrastructure and another set of competitors for SaaS. Customers need to extend these SaaS applications. We're by far and away the leader in ERP. People want to build data warehouses. They want to integrate with their on-premise systems.
They want to build all new custom systems associated with the ERP system. You got a choice. You can do it on Force.com. Well, they don't actually even have the ERP system, but the SaaS supplier, to be successful, is going to have to provide a facility for extending those applications. That, by the way, is called an infrastructure cloud. Workday doesn't have anything. How is Workday going to go in the ERP business, where they're not doing very well, without any ability to extend their applications? I have no idea how they do it because all of our ERP customers, once you get to a decent-sized ERP customer, want to build data warehouses around their applications.
They want to mash up on-premise data and SaaS data, then they want to do analyses, and they want to build custom stuff, and they want to do a lot of things. Well, Workday's answer is that we don't have an answer. We haven't thought about that. The market, people just draw this line between these two markets. We think a lot of people are going to come to our infrastructure cloud starting with SaaS, starting with ERP and supply chain and manufacturing, then build out from there a lot of custom stuff in our cloud. We think these are related businesses, not unrelated businesses, even though our competitors in each space are very different. Back to this. Our Gen 2 Cloud, when I say Gen 2 Cloud, that includes everything, SaaS, infrastructure everything. The network, the cloud control computers, the application computers.
There's this thing called bare metal compute. Some people refer to our Gen 2 Cloud as our bare metal cloud. bare metal compute is one of the options in our Gen 2 Cloud. If a customer is security conscious, they can choose not to share their computer, their customer computer with other customers, and they certainly not share it with us. When you buy bare metal compute, there's no Oracle code in that computer. There is no other customer code in that computer. You as a customer, don't have to trust anyone. You get bare metal compute. Again, that's just an option that's a part of Oracle's Gen 2 infrastructure. Gen 2 infrastructure is a part of Oracle's Gen 2 Cloud. Next slide. I mentioned security, reliability. That was the driving force of why we rebuilt our cloud, started from scratch.
We think we have a huge differentiator in the cloud business that is particularly attractive to our customers who use the Oracle Database. I think most enterprise information, most government agencies' information, the majority of it is in an Oracle Database. We want to make it easy for them to lift those databases up and those applications up and bring it to our cloud. We want to make that very easy to do. We have three design goals. One is security, reliability. It's got to have that. Second design goal is we've got to protect our customers' existing investment in Oracle Databases and applications on those databases. What does that mean? They have to be able to lift those applications up intact, those databases up intact, and move them to the cloud. Well, why would they bother to do that?
Because of the third design goal, which is there are huge economic benefits of doing that. By the way, it should run faster. It should be more reliable. It should have built-in disaster recovery, all of those features. It should just be much cheaper to do it that way because we have all these economies of scale, the things you're going to expect from the cloud. Take your existing applications, lift, shift them, press a button. They run faster, they run more reliably, built-in disaster recovery, and your bill is much lower than when you were running it on-premise. Those were our design goals. The driving force for the redo was the current cloud systems are penetrable. We think these cyber attacks are going to get worse, not better. We also think we have huge performance advantage.
I'm not going to go into it right now why, we think our Gen 2 Cloud is much faster than our Gen 1 Cloud. Should be. We're much smarter when we built it. It's a much flatter network. I'm not going to go into all of this stuff. I will show you some benchmarks. Next slide. This is kind of making fun of Remember where the cloud came from. It was Netflix. It was Angry Birds. It's great. I think Amazon's a fabulous company, by the way. Some companies I admire, some companies I don't. Amazon is an admirable company. They're amazing. They did a great job, and they did a great job innovating. The idea of picking this idea of renting compute and renting storage and renting networks was very clever.
I thought we were very clever when we invented, and we did invent SaaS, starting with NetSuite. Took us a while to get all of our Fusion apps up and running, but we're way ahead of everybody in SaaS. If you look at the breadth of applications we have. We have all the applications Salesforce has and all the applications Workday has, and then a lot of the applications no one has. Our SaaS portfolio is enormous, and we have lots and lots of integrated pieces because we started a long time ago. We are the creators of SaaS. Amazon, they're the creators of infrastructure. When they did it, their design goals and their design points and their customers were very different than a typical Oracle customer, a big bank, a big telco, people who are running their financials on this thing.
It's just a very different profile. Amazon's come a long way. They've made a lot of progress, but they have not changed their fundamental architecture, and they're going to have to. Everyone's going to have to. Next slide. The Gen 2 Cloud architecture, again, we built the Autonomous Database on Gen 2 Cloud. It is the foundation for Gen 2 Cloud. It's the foundation for all of our Fusion SaaS apps. NetSuite is being moved to Gen 2 Cloud. They're very excited about it because it's cheaper, it's faster, it's more reliable, more available. All of our SaaS applications are moving to Gen 2 Cloud. All of our customers are moving to Gen 2 Cloud. Our Cloud@Customer, which we've been selling, the main reason people are buying our Cloud@Customer is to get the Oracle Exadata Database Service kind of on-premise in a cloud form. That's the major reason they're consuming that.
Now with our Gen 2 Cloud@Customer, they're going to be able to press a button, and they get the full Autonomous Database@Customer. The full Autonomous Database cloud. They get all of the infrastructure, everything necessary that that cloud works, runs the Autonomous Database. We automatically back it up. We automatically patch it. We automatically look for threats. Threats are patch when it's running. They get full autonomy behind their firewall. On the cloud, because a lot of our customers are big banks, phone companies, who don't want to move to the public cloud, don't trust the public cloud. Regulators won't let them move to the public cloud. A lot of our customers, the intelligence agencies, are not going to move to the public cloud. Not that we won't build a cloud specifically for them.
That's what we have to do for some of these customers, is deliver Cloud@Customer, a region of our cloud built specifically for a customer. We can do this very inexpensively now for all of our customers with delivering Autonomous Database at Customer, which is a huge opportunity for us. Next slide. Anything interesting there? Yeah. It's more capable, more secure. You've heard all of this. We have SLAs. We're the only ones that give you uptime SLAs without a lot of caveats. Amazon has an uptime SLA, but it doesn't include patching, doesn't include software failures, doesn't include hardware failures. Read their SLAs. They say, "We guarantee this much uptime, but we don't include any of this stuff." We include all of that. The Oracle Database is down a couple of minutes a month. We guarantee it's down a couple of minutes a month.
We keep running if there's a database software failure. We keep running if there's a database hardware failure. We don't fail for any of that. We keep running if the whole data center is blacked out. Next slide. We're much faster. Yeah. This is funny. We're much faster than Amazon. A lot of people say, "Yeah, Oracle, you're always bragging about performance, but Amazon's so cheap." When we charge by the minute and they charge by the minute, and we're 10 times faster, if we have the same price per minute, then they're 10 times more expensive than we are. We have a pretty aggressive pricing strategy. Our performance advantage translates into incredible economic advantages if you compare our cloud to Amazon. Next slide. Next slide. Okay. This is just basic infrastructure. Our basic compute is 50% faster than Amazon.
Our block storage is six times faster than Amazon. Our network is twice as fast as Amazon. That's not the most interesting part. Next slide. The economic advantages are much bigger than the performance advantages. Amazon is three times more expensive for compute, triple what we are. They're 30 times more expensive for block storage and 10 times more expensive for the network. Amazon is coming back with a lot of comments about my presentation, but they're not saying, "You know this slide? This is wrong." They say, "Oh, that's Oracle saying a lot of crazy stuff." None of which is true. Can you be a little more specific than that? This is not subtle. Why don't you just take this slide and say, "They're lying about that. We're cheaper than they are"? We're not lying about any of this stuff.
We're much faster, and we have aggressive prices. Next slide. This is block storage, where we're 6 times faster, but the price difference is crazy. It's absolutely crazy. Next slide. OCI and big data workloads. This is a TeraSort. Again, 3 times faster, 8 times less expensive. I should really have 3 times faster, X times cheaper, so you don't have to do the calculations in your head. Anyway, it's usually 3 times faster, 4 times faster, 10 times cheaper. These are just huge advantages versus Amazon. Next slide. Why are we so much faster? We have a much more modern system than they do. The great thing about starting over and saying, "Okay, let's do this one more time. Let's redesign this. We're now smarter about this" our network's totally different than theirs. We have an RDMA network. That's remote data memory access.
Very simply, it means one computer can take a bit of data and move it to the other computer without those computers really talking to each other. The operating systems aren't talking to each other. There's no interrupt. There's no time lag. It's just suddenly someone just reached into your computer and just stuck data in that memory, and then taps you on the shoulder when they leave. They don't have an RDMA network. We have huge broadband network with very low, microsecond, between one and two microsecond latency. They don't have any of this stuff. Next slide. By the way, we needed it. We had no choice. We had to build it because that's how Exadata works. The problem was we weren't starting with, I don't know.
Amazon, great thing, you're the innovator, and you say, "I wonder who the first customer is going to be." I don't know if they ever would've anticipated Netflix, Angry Birds, gamers, and all of this stuff. We knew who our customers were. They were our existing customers who wanted to move to the cloud. We had to make Exadata work in the cloud. You can't make Exadata work without an RDMA network. We have to build all this stuff in. These are just a bunch of engineering simulations. Again, we're much faster than they are and much cheaper, it's the same story. Next slide. This is, again, also computational fluid dynamics, it's over and over again. If Amazon wants to say, "Ah, they're just making this stuff up," pick any slide. Every slide is backed with a benchmark that's published with all of the details.
You can go to our website, get all of these details that we are this much faster and this much cheaper than Amazon and with our Gen 2 Cloud. Next slide. Of course, the biggest thing is security. We have all of this security stuff. I'm not going to go into all of it. The big thing was re-architecting these things that we protect the perimeter of the cloud, we protect user zones. Our cloud control code is in separate computers than theirs, and tons and tons of AI, tons and tons of machine learning to detect threats and kill them while the system still runs. Next slide. We have a key management service. Everything's encrypted, we have this unique ability. Our customers don't trust us with the keys. It's very interesting.
Your bank or somebody said, "I don't trust Oracle to manage my keys. I want to manage the keys myself. I want everything encrypted in the cloud, but I just don't trust these guys." It's not a problem. We let you do your own key management. We actually have hardware that specifically allows you to do that. You don't have to trust us. Amazon can see all your data. A lot of people don't care. Again, a bunch of people don't care. Workday can see all your data. Salesforce is a little bit different. I actually don't know what the situation is with Salesforce because I don't know what Oracle options they're using. Salesforce is all built on top of Oracle, and I don't know, we have this facility feature called Data Vault, which if they turned it on, means they can't look at your data.
If they didn't turn it on, they can look at your data. The truth is, I don't know if that's what the situation is with Salesforce. Workday, guarantee you, they can look at all your accounting data, all your HR data. That's just not acceptable to a lot of people. Next slide. Yeah, just more security stuff. Application firewalls. We paid a lot of attention to security. Security is in our very first customer is the Central Intelligence Agency. Our second customer is the National Security Agency, blah, blah. These people care about that. Next slide. We're rolling out our Gen 2 data centers all over the world. In fact, since this slide was created, we added UAE. We have two separate data centers in the Middle East. Other candidates are coming up.
We're going to be building these Gen 2 data centers all over the place. Next slide. We'll build a Gen 2 data center for customers. If you're a large security-conscious customer, if you're a big bank, if you're a government intelligence agency, we'll build you a Gen 2 data center that is all your own. That's all your own. Next slide. The Gen 2 stuff, the Gen 2 infrastructure, bare metal compute, all that stuff's available right now. Our Fusion SaaS applications are already running on our Gen 2 data center. Other SaaS applications are being moved. We are moving our Cloud@Customer to Gen 2, I hope before the summer, I hope as early as January. Press a button and get Gen 2. I'm not sure when we finish. It's an engineering project.
All of our customers will be able to get Autonomous Database in cloud form on their premise behind their firewall. The people who have already bought the Exadata on-premise will be able to press a button and get Autonomous Database. Next slide. Okay, I'm going to go through this very fast. In fact, I think I've Next slide. How am I doing for time? All right, I promise I'm going to go very fast. All right, lots of automation. It's a self-driving database. There are no database administrators. You press a button, it gets compute, it gets storage, it gets network. If it needs more compute, it gets more compute. Automatically backs itself up. Data center fails, keeps running. Computer fails, keeps running. Never breaks. Next slide. Again, it's serverless. When you're not running the Autonomous Database, you don't get a bill. It's not like Redshift.
Serverless means when you're not running, you're not paying for any compute. Your compute cost goes to zero. It's very economical. It's very, very economical versus other database cloud services. If you're running on one computer and suddenly it's the month-end close and you need three more computers, it adds those three computers for four hours, then takes the three computers away. It's truly elastic while running. Amazon can't do any of this stuff. Amazon can go from one core running Redshift to two cores running Redshift, or one core running Aurora to two cores, as long as those cores in Aurora. By the way, Amazon Aurora, which is Amazon's OLTP system, transaction processing system, is really just our MySQL open-source system, renamed Aurora. We built it, Amazon didn't. That's their transaction processing system, and it runs on only one computer. There's no elasticity.
There's no ability to run a second computer. If that computer breaks, by the way, you're down. This is another claim I'm going to make. Aurora, not built by Amazon, built by us, called MySQL. Amazon just gave it a new name. It's also true of Redshift, right? Amazon didn't build Redshift. That's not what they do. These are just open-source pieces that they use to build their cloud. They did a great job. They did a great job of making those pieces available in a coherent, deliver them in a coherent cloud. I give them a lot of credit. They don't build databases. Amazon still runs all of Amazon on Oracle Database. They are trying, they promise, because they don't like me reminding them of that publicly. They said they're trying to get off Oracle by 2020.
That's very interesting because SAP's been trying to get off Oracle for 10 years, and SAP has its own database called HANA. SAP still runs SuccessFactors on Oracle. They still run Concur on Oracle. They still run Ariba, and they've had Ariba for more than a decade. They run Ariba on Oracle. The EU did a study of how many SAP customers in Europe, big customers in Europe, used Oracle, and they looked at their 100 top customers. We didn't get all 100. We only got 99. Turns out that Nestlé used IBM Db2. SAP has been trying to get off Oracle for a very long time. It's not easy. Amazon just moved a bunch of their warehouses to Aurora and shut them down cold. You can read the Amazon documents. We can send you copies of the Amazon documents. Aurora is our other database.
It's our low-end database. It's our database. It's not theirs. It's open source. Anyone can use it for nothing. They moved to Aurora. Their warehouses were shut down. Next slide. You don't read about Oracle Database being hacked and data being stolen. Doesn't happen. Next slide. By being fully autonomous, you eliminate human error. Most airplanes crash because of pilot error. Most cars crash because of driver errors. Computers are better at this than we are. It's much cheaper to have autonomous systems. The Oracle Database not only is much faster than Amazon, but there's no human labor, it's much cheaper to have your data in an Oracle Database than have it in an Amazon database. It's much more reliable because there are no human errors. Plus, if something breaks, we tolerate the failure. Their systems are not fault-tolerant. Next slide. Next slide.
We've been doing this for a long time. We've been adding a lot of automation to our database software. Next slide. That's not enough. Next slide. To build a truly Oracle Autonomous Database, you have to automate the database software. You have to automate the underlying infrastructure completely. If I say a data center fails, and your application keeps running uninterrupted, obviously, something has to happen that's beyond just the database software. We have to automate the database software, automate the infrastructure, and automate the inter-data center connections and all of that. Next slide. That's what we did. That's what the Oracle Autonomous Database is. It's a combination of infrastructure automation, database automation, and data center automation all rolled into one to get a full autonomous system. Next slide. I'm not sure I'm going to do this. Next slide. Next slide. Next slide. Wish I could just click.
Next slide. That's fine. Okay. This is interesting because we have one database for transaction processing and data warehouses. Amazon has one database for transaction processing, Amazon Aurora, and they have another database for query processing, Amazon Redshift. We have one database. You will see why their approach is problematic. Next slide. By the way, all of our stuff is autonomous. We came up with Autonomous Data Warehouse about a year ago, and about six months ago, we announced Autonomous Transaction Processing. Next slide. This is very interesting. When we took highly tuned systems, from a stock exchange, a manufacturer, and a bank, threw away all their tuning and let the system tune itself, computers were better than the experts. These are all real customers. Next slide.
We even took NetSuite, which is tuned by lots of smart people who work for us, and just threw away all their stuff, and the system tuned, did a better job than they did, though the difference was much smaller than a typical data warehouse. Next slide. All right. Next slide. Just like we have bare metal for general compute and infrastructure, we also have bare metal for Oracle Autonomous Database. So you can say, "I want that Exadata machine," and all Oracle Autonomous Database runs on this Exadata machine. "I want that Exadata machine all to myself. I don't trust anybody. I don't want anyone else's code, anyone else's data. I want only my stuff. I want you to give me a completely isolated system." We can do that. Sharing tends to be cheaper than having dedicated hardware, but dedicated hardware gives you more security.
We've introduced that option to our customers. Next slide. Again, I mentioned this. Press a button, you get full Autonomous Transaction Processing, data warehousing at customer, and Gen 2 Cloud@Customer coming out early next calendar year. Next slide. Next slide. Again, the Autonomous Data Warehouse. Next slide. Again, we guarantee that your Amazon bill will go down by half. If you're running on Amazon Redshift, whatever you're running on, if you bring that same workload to Oracle, we will guarantee that our bill to you, that bill will go down by half. That's not even the big savings. The big savings is there's no human labor. There are no human beings running our system. Our system's autonomous. You need human beings running their system. Plus, our system doesn't break, and it's secure and has all these other advantages.
Just your Amazon bill, we guarantee, will go down by half. Next slide. We can guarantee that because we're so much faster than Amazon systems. Next slide. Next slide. All right. Oh, we don't have the benchmark slides in here? Oh, darn. This is a summary of the benchmark slides. The first one was a data warehouse application. You can run it. We have the workloads. It's all published, all the details are published. We're 9 times faster, 8 times cheaper than Amazon. Same exact data warehouse, Oracle versus Amazon Redshift. We tested this a year ago. They haven't gotten any better. They're not in the database business. On our new transaction processing system versus Amazon Aurora, standard benchmark, standard transaction processing benchmark. Okay, we're 11 times faster, 8 times cheaper. This is the one that's interesting.
What if you put queries and transaction processing into the same workload? What if you test Amazon Aurora? That means Amazon Redshift doesn't do transaction processing. The only option is Amazon Aurora. What happens if you have a mixed workload of queries and transaction processing? We're 100 times faster than there. They don't have a database that can do normal things. A normal thing is the combination of transactions and queries. We're 100 times faster than there. You think they would say, "Well, this is really a lie. This is a big-time lie. Let me just point out how wrong this is and explain to you why they're just making this crap up." Not a word. This is all published. It's all published. Not a word. Okay. I'll tell you, they have one great database that runs on Amazon. The next one.
Oracle, when they run the Oracle Database on the Amazon cloud, we're only 3 times faster. That's the best they can do when the customer brings their Oracle Database. By the way, if there's a bug or you're patching, or there's a hardware failure, then we're infinitely faster because they're down. They're down a lot. When they're down, you're still paying for the computer. It's not serverless. We're infinitely faster and infinitely cheaper in that case. Next slide. Okay, this is just a summary. The Gen 2 Cloud. We had to add a new collection of computers to the Gen 2 Cloud to protect the perimeter of the cloud and protect the perimeter of every user's zone. Autonomous robots find threats and kill them while the system is still running. The Oracle Autonomous Database, no human labor, no human error, never goes down.
If you want a system that's highly secure and never goes down, you have to be willing to pay a lot less. I'm going to stop right there. Okay? All right. Yes.
Please.
45 minutes. I think they said 45-minute presentation, They're 40 minutes into the presentation.
40 minutes, Larry. 40 minutes. Start with Michael.
Thanks, Larry, very much for all that. Is the goal and the focus of Oracle Cloud Infrastructure still primarily to run Oracle workloads? How much are you trying to optimize non-Oracle workloads? If not, can you get to the same scale as your other competitors?
Well, I think you saw earlier in my slides that we compared TeraSort, big data, high-performance computing, scientific computing. Our goal is to pursue all workloads. However, we think we have a huge built-in advantage. Interesting question. In corporate America, how much of the data in, I don't know, pick AT&T, Bank of America, how much of their data do you think is stored in an Oracle Database? Most of it. I think most of the world's data is in an Oracle Database. Most of the world's high-value data. I'm not talking about cat videos. Because cat videos have two problems. One is, it's a lot of data. I love cat videos, by the way. It's a lot of data, but it's not really high-value data. Bank of America probably doesn't have a huge number, AT&T doesn't have a lot of cat videos.
If you get rid of the silly things like movies and all the data that's created on your iPhone, it's great to store all that someplace. If you look at corporate information, government information, airline reservation systems, banking demand deposit accounting systems, loan origination systems, switching systems in phone companies, inventory systems and personnel systems in militaries. If you look at those applications, what percentage of that data do you think is stored in an Oracle Database? I'm asking you, what do you think? Most of it. If I were to say we're just going after Oracle workloads, if we got the Oracle workloads in our cloud, how big would our cloud be? Anyone want to guess? Seriously, I'd love for someone just- $100 billion. $100 billion? Yeah, I think certainly north of $100 billion, right? $100 billion would be a very safe conservative estimate.
It's a lot. We're greedy, and when we make the infrastructure fast for Oracle workloads, kind of accidentally, we make it fast for other workloads. We don't know how to make it fast for Oracle workloads. For example, CERN is a huge customer of ours, and they store vast amounts of data in the Oracle Database. It's obvious, and they do a lot of high-performance computing. As we built our very fast network, our RDMA network, which we had to do for Exadata, that was perfect for high-performance computing. Amazon doesn't have that. As we make things run fast in our cloud, we accidentally optimize non-Oracle workloads, even if we're not trying to do that. If we were to get Oracle workloads, it's actually way more than $100 billion. It's way north of that.
I think if we were just to get the Exadata workloads, probably be $100 billion. If we were to get kind of all of the Oracle workloads, it's crazy large because it's most of the world's data, most of the world's high-value data. I believe my friends would say, "Well, that's a big business opportunity." We're pretty optimistic about our ability to get a significant percentage of the Oracle workloads. The next thing is I also think in applications, if you look at the last application war, which was between us and SAP, and you kind of forget, and SAP won. We were second. I think they had about a 25% share, and I don't know how you figure our share at 12%, 13% or something like that. Maybe. I've seen our share up at 16%, but I'm not sure it was that high.
Between the two of us, we had, what, 40% of the market? There's another 60% of the ERP market that neither one of us have. I don't think that's going to happen within the new generation of cloud ERP. I think it's going to be like everything else. There's Google Search, and there are other people who try to be in the search business. Let's ignore China just for a minute. It's like another planet. It's against the law for Google to do search in China, unless Google is willing to do some unnatural acts, and maybe they are. If you ignore China, the network effect is very, very powerful in the cloud, and you suddenly get the network effect for ERP. You get the network effect for HCM and all of these other things.
Once you get ERP customers in the cloud, they use your infrastructure to expand and build data warehouses and do all of these other things. The network effect is not only horizontal, the more ERP customers you have, the more ERP customers you'll get. They start transacting with each other via the ERP systems. Business A buys something from business B. That's just an Oracle Fusion ERP purchase going from one ERP system to the other ERP system. It tells you when available promise.
There's the horizontal expansion network effect, there's the vertical expansion, the vertical integration, the fact that we have infrastructure, we have a Gen 2 Cloud underneath our SaaS applications, which really lets them build out and keep in the same data center their data warehouses, their ERP data, their HCM data, their field service data, keep all of that stuff together, is a very interesting story. It's all secure, it's all reliable. It's all built on this highly reliable infrastructure, highly secure infrastructure. We think that's going to be Well, it's kind of exotic right now, we certainly spent a lot of money and put a lot of effort saying, "Okay, we're going to redo what we have." We think that puts us in a very good position going forward. Yes, sir. Yes. Yep.
Hi, Larry. Keith Bachman from Bank of Montreal. I wanted to get your thoughts on what you think the timeframe associated with autonomous database will be for your customer set. Secondarily is, how do you think about Mark and Safra spent some time talking about how applications may indeed inflect up here in terms of growth rates. How are you thinking about autonomous database ability to help the growth of platform and infrastructure category as you look out?
Yeah.
Thank you.
Truth is, I don't know. That's the real answer. How fast it grows, I just know that no one has anything like the autonomous data. There are a huge number of people who use Oracle, and they'll get huge benefits from using Autonomous Database. Here's the catch. Most of our customers are on-premise, and Autonomous Database is a cloud product. It's a really interesting problem. To adopt Autonomous Database, you have to move toward the cloud. The good news about just focusing on Autonomous Database, the good news is this new Gen 2 Autonomous Database Cloud@Customer. For the first time, sometime early next year, we're going to be able to put Autonomous Database on-premise, albeit it's called Cloud@Customer. Just like JEDI, Amazon's JEDI bid, that's Cloud@Customer. That's not really Amazon's public cloud.
CIA is not using Amazon public cloud. We're bidding on that. That's not a public cloud. It's just a big data center going into a customer. We have ambitions of building big data centers, not only for the government, not only for the security agency, which is our heritage, but big data centers, Gen 2 data centers for our biggest telco customers. Is our plan to do that. I think some of our smaller customers will adopt. The majority of our customers, I think, will adopt Autonomous Database in our public cloud. I think with Gen 2, we've done a bunch of things on Gen 2. Our biggest focus in Gen 2, to be completely forthcoming, is just getting the usability down where it's really easy to get to Autonomous Database.
We had all of these features and functions, to tell you the truth, the UI, just navigating our cloud wasn't that easy. Amazon was way ahead in terms of ease of use of the cloud. We think right now, literally right now, our cloud is as easy to use as Amazon, and we have ambitions to be better than, easier to use. It should be easier to use because Autonomous Database is much easier to use than their databases. We've got to make it very easy for our customers to get to Autonomous Database in the public cloud. We think we're kind of there now. We think the uptake in 2019, we're going to see a major inflection point in public cloud in 2019, also for Autonomous Database, also Cloud@Customer. I can't tell you the slope of that curve. I'd love to.
No one wants to know more than me. I think another way to ask your question is, "Hey, great presentation. When are we going to see it in the numbers?" Next year for sure, and I just can't tell you the slope of the curve. It's a very big deal for us. There are two strategic businesses we have, Database and ERP. ERP is the biggest SaaS application segment, and we are right now dominant in the cloud in ERP. SAP is dominant on-premise in ERP. We're number two. In the cloud, I think the biggest piece of infrastructure is going to be database because it's called the information age, not the compute age, not even the movies on demand age. It's called the information age, and I think the database is a very big deal. We're the dominant provider of Database on-premise.
The question is moving all of that to the cloud. We decided to bite the bullet, build Gen 2 Cloud, and then start encouraging our customers to move off-premise into our public cloud using Autonomous Database because now the public cloud is secure and reliable and all of those great things. I expect an inflection point next year, but I can't tell you what month. If the question is when are you going to start seeing it in the numbers, I think sometime in 2019, it's going to show up in the numbers. At some point, it's going to have huge impact. Yes, sir.
To your point about the presentation, how it sounds compelling, how tied is the movement of the database to the cloud, to the Autonomous Database, to customers moving their applications to the cloud, right? The obvious answer would be that they're tied very intimately. Then secondarily would be, I just recall one example of an ERP implementation. I think it was Pepsi came out with a press release a couple years ago, and they said they had just completed their global ERP migration, and it must have been like, I don't know, 15 years. I don't know how many hundreds of millions. Customers have a depreciation cycle on this stuff, right? How tied is the momentum of your installed base moving to the cloud with their ERP and application systems to-
This is a SaaS-specific question?
Well, I don't know. You tell me.
Okay.
I was asking my host.
Okay. Let me talk. There's two kinds of migration to the cloud. There's the migration to the cloud where you take your existing application and your existing database, you lift it up and you just drop it down to the cloud. As a result of doing that, it runs faster, it's more secure, it's more reliable, and maybe the most important thing, it's just much cheaper to run. That customers can do in Gen 2 right now. That's a rather straightforward technical process. It's not a terribly difficult technical process. Remember, the second design point of our Gen 2 Cloud was easily lift and shift database and related applications. Lift the data up out of on-prem, drop it down to our cloud. That's pretty easy.
What I gleaned from your question was, people who have spent a lot of money implementing ERP, SAP ERP, some of these implementations cost $1 billion. SAP is actually going back to their customer base right now and saying, "Hey, we got the new generation of SAP. The current generation you're on expires in 2025. You need to start planning now to move to the new generation of SAP," which is really cool because it uses all the SAP tried-and-true code. There's no risk, really, except it all runs on the HANA database. You can get rid of that Oracle database, you got to move again. I know that you've heavily modified that version of SAP. That's why it cost $1 billion. You can't implement a standard application for $1 billion. It's not possible.
You have to hire a lot of people to write a lot of code and make a lot of modifications and a lot of extensions to the code. You're going to have to do that again as you migrate to the new version of S/4HANA in the cloud. S/4HANA in the cloud has nothing to do with the cloud. S/4HANA in the cloud is the old SAP code and running on the new SAP database that has exactly zero to do with cloud. It's the same on-premise code that they wrote 35 years ago. There's almost no new code. There's a slightly new UI on it, anyway. SAP is now telling their customers they're going to have to move. Thank you, God. They've actually submitted billion-dollar proposals to some of their customers to get them going on the next-gen S/4HANA in the cloud.
How much will that cost? Actually, for you, not $1 billion. Cost you $1 billion last time. This time, only $900 million. That's a real number from a real customer. SAP proposed a $900 million move to S/4HANA in the cloud. SAP is doing us, I think, a very big favor by pushing their customers off the Oracle Database onto HANA, because the new S/4 only runs on HANA. It doesn't run on any other databases. To get the benefits of HANA, you got to be willing to pay $900 million. It's a lot of money for a database that loses benchmarks to Oracle by 10 to 1 and doesn't have any of the reliability features or any of that other stuff that I talked about. There's a huge opportunity in the SAP install base, and it's happening now because SAP's end-of-lifing their current product.
That's a big opportunity in SaaS. There are other opportunities. That's SAP, which is a real company with real engineers. I think some of these companies are going to pay the $1 billion. They're going to say, "Hey, I'm screwed. SAP's got me. I can't take the risk of moving something else. I'll pay the $1 billion." A lot of rich companies are not happy with this, and I think we can get a lot of those guys moving because we can move those companies off to the cloud or our ERP in the cloud for 1/10 the cost. Seriously, 1/10 the cost and much faster, no risk. We are moving people. I'm not going to go down a list, but we're moving SAP customers, big SAP customers, to our cloud. Big SAP customers. Companies you've heard of, guaranteed.
We're doing that now, it was all kind of provoked by SAP saying, "Hey, I need another $1 billion." There are these other companies like Lawson, also AKA Infor. That Infor bought all these companies like Lawson and in the healthcare industry. What are you going to do? Wait for the next generation of Infor? We are moving a lot of this other 60% that wasn't Oracle or SAP, these guys got no place to go. These guys are desperate, and they're moving at a very rapid rate. I can mention Cleveland Clinic, I can mention a whole bunch of companies, and healthcare companies that are moving off of Lawson onto our cloud ERP. They're moving as a group.
They all talk to each other. Kind of the whole industry is going to move off of Lawson Software and onto, off of Infor and onto our ERP in the cloud in the healthcare industry. There are other industries I can talk about. Our own customers, I saved you. I used up my whole 45 minutes on infrastructure and database. Our own applications, Fusion applications, have all this automation, a cool new voice interface. Our customers, our PeopleSoft ERP customers, our E-Business Suite ERP customers, our JD Edwards ERP customers, they now are motivated to migrate from their current ERP system to Fusion. We've actually built a system, an automated system, a machine learning system that will take an E-Business Suite customer and do a semi-automated, it's not autonomous, a semi-automated upgrade to take you from your on-premise system to the cloud. A quick summary.
There's two kinds of moving to the cloud. There's one, like in ERP, where you throw away your current ERP system, like E-Business Suite, and you move to Fusion, or you throw away SAP and you move to Fusion, or you throw away Lawson Software and you move to Fusion. That's a project. There are huge benefits because you get this very modern cloud-based system. There are other moves to the cloud that are simply lifting what you have. Don't change it. Lift the Oracle Database, lift the application, and take that and just drop that into our cloud. That's a much simpler process. We're in the middle of both of these things.
In a way, you would think the harder of the two would be winning in SaaS, because you got to have people basically throw away their current ERP system, throw away their current HCM system, and move over and do a re-implementation with your cloud system, where it's not a lift and shift. It's a re-implementation. We're doing very well there. We started a long time ago. We are on our way to being by far the largest SaaS supplier in the world. We're going to go from number two in applications behind SAP, I believe, to an overwhelming number one in SaaS. It's what I believe. Okay? I might be crazy. The easier of the two things, I said Oracle has two strategic businesses, ERP, SaaS, get all of those applications over to the cloud.
There, the prize is to go from number two to number one. It's a big business. It's a $20 billion-plus business, right? The bigger prize, in a way, is much simpler to win, because all we need is for our customers to lift their existing Oracle Databases and their existing applications up, and then just drop them into our cloud. Albeit building that cloud was very building a secure cloud, making this thing reliable, turned out to be really hard to do and took much more time than we ever thought it would take. We're there. Now this lift and shift could happen relatively rapidly. That's why the earlier question, when do we see the inflection point on Autonomous Database people taking their existing workloads and moving them over to the cloud?
I think now that Gen 2's there, Oracle Autonomous Database for transaction processing and data warehouse is there. All the pieces are there. We just have to help our customers get a few workloads moved, and after they get the first few moved, they should be very motivated to move the remainder of their workloads. Yes, sir.
Hi, Larry. Anurag Rana from Bloomberg Intelligence. As I look at the hyperscale cloud providers right now, Microsoft has a good on-premise product base, good public cloud. Amazon didn't have a good on-premise base, so they went and partnered with VMware. Google doesn't have a good on-premise product base. Why not partner with them? It'll help you and it'll help them.
Say again?
Why not partner with Google? It'll help them and it'll help you.
Why not partner with Google? We're suing Google. We're suing Google. They stole Java from us and built the Java telephone, which they called Android. They were found in federal court to have stolen, infringed on all of our copyrights. Now why wouldn't we partner with them? The next phase of the trial is to come up with damages, see how much money they owe us for stealing all of Java. Why wouldn't we partner with Google? Google's cloud is not like our second generation. A, it's very hard for two companies. Why doesn't Ford partner with Tesla? They're competitors. We're a competitor of Google. Ford doesn't partner with Tesla because Ford was trying to compete with Tesla. We don't partner with Google because we're trying to compete with Google.
Google, I don't think is going to be a very good partner because they built the wrong cloud. They built a Gen 1 Cloud with all of these problems, and they can't run our Exadata workloads. They have not got an RDMA network. They have not separated their cloud control code from their application user code. They haven't done any of that stuff. I guess those are the two reasons that we're not partnering with Google. They're thieves, and their technology is not that good. If it was search, we'd love to partner with them in search. Yes, sir.
Thank you. This is Keith Weiss from Morgan Stanley. Thanks for taking the time to talk to us. In talking to a lot of partners this week, including some of your biggest size, you definitely get the sense that Gen 2 architecture is really ready to turn, and they're really excited about the opportunity within the install base. I think some of their benchmarks might be even better than yours that they've run internally. Mark gave us a really good sort of benchmark of how to think about application customers coming over to the cloud in terms of that 3x. Can you give us kind of like an idea of how we should think about an existing on-prem database customer when they come in to use the cloud? What does that look like?
I think that's the right question if you're a betting man, right? If you have anything to do with trying to predict markets and the investment community and doing this stuff, that's the question. I think I can tell you what the end state looks at, but it's so crazy. What you really want to know, if I'm an investor, by the way, my second largest investment, I will disclose it now. I'm not sure how many people know. I'm very close friends with Elon Musk, and I'm a big investor in Tesla. Tesla had a good day. I think Tesla has a lot of upside. Do you know the most popular car by revenue in the U.S. over the last three months? The number 1 car in the U.S. by revenue? The Tesla Model 3 outsold in revenue.
Forget about Ford and GM. They weren't even close. Outsold Honda and Toyota. I loved all the articles about Elon doesn't know what he's doing, the pictures of him smoking dope. The Wall Street Journal writing all these articles. He's going to have to go out for money. I'm really smart. I work for The Wall Street Journal. I know a lot about writings for The Wall Street Journal. This is not going to endear us to The Wall Street Journal, though I read it every day. I do. It's my favorite newspaper. I'll be honest, it's my favorite newspaper. Yeah, that's my favorite newspaper. More than The Economist now. Then people, "Oh, they're just going to have to go out for more money. This is all nonsense." I said, "Who are you? This guy's landing rockets. He's landing rockets on robot drone rafts in the ocean.
You're saying he doesn't know what he's doing. Well, who else is landing rockets? You ever land a rocket on a robot drone? Who are you? Okay, you're telling me he's an idiot. You're telling me he's an idiot. I just want to know who you are so I know why should I believe you as opposed to my friend Elon, who and we're out here watching this rocket land, and which I think is really cool. You're there in front of your Apple Macintosh and typing up an article saying Elon's an idiot." Okay. All right. Had to get there. Tesla had a good day. Picking what quarter. Let me tell you something about Tesla. Elon says, "We're going to get this done.
We'll ramp up the Model 3s by Q3." He's got the whole Tesla team excited, and we're trying to get this done. We missed Q3. God damn it. We're going to Q4 for sure. Motivating the team. Setting tough goals. It turned out it's very hard to predict what quarter. There was some relation to your question, my little riff here. You're saying, okay, what quarter does Oracle just kind of surprise you? Where'd this revenue come from? We weren't expecting. This point just kind of snuck onto our chart from nowhere. You quote your revenue is kind of flat, and actually it's not flat. Some things are growing very rapidly and other things are shrinking, and we're kind of managing to flat, right?
The only way it could be flat for this length of time is if we are kind of de-emphasizing businesses we don't care about, growing businesses we do care about, and going through this transition in a kind of a market acceptable way. That's what we've been doing, trying to completely. Underneath the covers, underneath the revenue, and you look at this business, some of the businesses are exploding. Some of the businesses we just don't care about. We want to shrink them. A lot of those businesses we don't care about. Anyway, there's no way to. Until we get the first data, a few data points on Autonomous Database to see what the adoption rate is and kind of start to measure the slope of that curve, there's no real way for me to tell you. I mean, is it First calendar quarter in 2019?
Is it second calendar quarter in 2019? I don't know. I do know like Elon knows, this is a pretty good product. We got to get it to the market, and we're going to sell a lot of this stuff. I just can't tell you exactly what quarter. Now, when it hits and you get a couple of data points, let's say it hits in Q2, and then you get another data point in Q3, you start to sense the slope in the curve. It's growing, and then all you guys can place your bets. I fully understand somebody who would say, "Well, I don't see it in the numbers." Hell of a, as I say, nice presentation. When is it going to show up in the numbers? I'm just going to wait till when it's going to show up in the numbers. You can do that.
That's the choice, right? I could have waited in terms of buying a lot of Tesla, wait for it until it just shows up in the numbers. Prove to me. I went, he'd drag you. We have dinner at the goddamn Tesla factory. Go to the Tesla factory. I went home. Elon slept there. Go to the Tesla factory, look at the Model 3 line, look at the Model 3, go to the Gigafactory, do all of this stuff, talk to the people, figure out good product. Really cool assembly line. Gigafactory, amazing. Are there problem? Yeah, there's huge problems with Panasonic because you're trying to get these batteries to work and your primary cell manufacturer, Panasonic, is having problems. You've got to work through that engineering. It's an engineering project. It's hard to get all these pieces together.
If you get all these pieces together, it's hard for other people to compete. I think at this point, with Autonomous Database and Gen 2 Cloud, we've got a really firm infrastructure foundation underneath our most strategic product, which is our database. Most of the world's data being stored in that, we think is going to migrate to a combination of our public cloud and Cloud@Customer. I don't know where else it's going to go. People have tried. Believe me, people have tried to move an Oracle workload to Redshift or Aurora. They've even tried to move it to Microsoft SQL Server. It doesn't work. We're much faster. We have many more features. There's a reason why we beat Microsoft in the database business. We beat IBM in the database business. We do have a majority of the database business.
Most of the world's valuable data is in an Oracle Database. Now we have an even better database in the cloud, and it's very easy for people to move. They're going to start in earnest in 2019, and where this ends up, again, low bid, $100 billion business. I can tell you some SaaS margins. When we get a new SaaS customer, we get another SaaS customer. What do you think the margin, someone gives us $5 million a year, nice size contract, $5 million a year to run ERP in the cloud. What percent of that $5 million do you think is margin? Incremental. 70. 70? Any other guesses? 90. 90. Maybe more, but certainly 90. It's a good business. It's a good business. Just like search. You get it all, it's a good business. Social, Facebook, you get it all, it's a good business.
ERP, you get it all, it's a good business. Database, you get it all, it's much bigger than any of those businesses I've just mentioned. Much bigger. Yes, sir.
Thanks. Hi, Larry. Brad Zelnick with Credit Suisse. Thanks for the time today. From this whole week, it's clear that your new Gen 2 Cloud is differentiated. It's a premium offering, and that's not a surprise given what we've come to expect from Oracle, whether it be database or applications.
I'm glad you said that because a lot of people have forgotten. We're actually a pretty good engineering company. No, I get it. This infrastructure stuff was way harder than we thought. When we did it the first time, it was not a very pleasant experience to kind of look at this like, well, I think I now know enough to know that copying Amazon was not a good idea. That doing what everyone else was doing was not good enough for our customers, so we had to do more engineering. We are pretty good at this. When the MIT Club meets, it meets at Oracle. A lot of talented people working on this problem. Anyway, sorry.
No worries. I didn't forget, but I think what is maybe a little bit more surprising are the price points. I don't think Oracle was ever known to have competed on price, whether it be in database or in applications. I just want to know your thinking behind the pricing strategy.
That is so-
When you have so much differentiation.
I will tell you a great story. Oracle's always had a very high list price. I remember I was at Davos, and I've gone through phases with Bill Gates, and Bill Gates and I are very good friends right now. In fact, I gave him $1 billion. Why would I give Bill Gates That's not true. I gave him $100 million. I didn't give him $1 billion, for his foundation on a particular project. I didn't give him $1 billion. $100 million was bad enough. He asked for $100 million, that's what he got. If he'd asked for more, maybe he would've gotten more.
I'm on this panel years ago with Bill, and this one, Bill and I were fighting and I said something about Microsoft, and he said, "Larry runs this company where people pay ridiculous amounts of money for a database that's only slightly better than ours. They pay 10 times more money for a database that's only slightly better than ours. I'll never understand that. Why would people do that? I guess it's because it's not their money." What I never told Bill was, it is true, our list price is 10 times more than SQL Server's. Microsoft sold all of their SQL Server pretty much through dealers. They didn't have direct salespeople. They have dealers, so their list price actually meant something. You got a discount off the list price, that's what you paid.
We sold our database for about half the cost of Microsoft SQL Server because we had these huge quantity discounts. We had discounts that were just 90%. In the early days, when we are competing with Microsoft, we would have these crazy discounts, all you can eat. You could not even compute the discount because they were unlimited licenses. We wanted to take down AT&T, we wanted to take down Bank of America. Now, we had a much better product than they had. We had a better product than IBM, but we were hyper price-- No one knows this. I am not sure I have ever told this story. No one knows this, that we were incredibly price-aggressive because it was a land rush, right? We knew we had to get AT&T. We knew we had to get all these big banks. We had to get everybody.
Once they were an Oracle customer, we would have an opportunity to sell them more stuff. During the highly competitive phases of the database business, when we had real competition from IBM and real competition from Microsoft, By the way, that eventually went away. By the time we were Oracle version 7, that competition went away. We were very price-aggressive. We undercut them all the time. Again, that was back in the day. That is the day, right? It is a land rush now for cloud. We have the opportunity to be very price-aggressive, where it is very hard for Amazon to meet us because a lot of Amazon's profit. It is a huge business for them, right? Right now, it is just found money for us. We already sell the database. We already get database license fees.
When you use our cloud, you are taking your existing Oracle license and bringing it to our cloud. We can afford to be much more aggressive on pricing than they can be for two reasons. One is we are trying to undercut them on a per CPU hour, just make sure our price is. Then we run 10 times faster than they do. It takes them 10 hours what takes us one hour to do, we are dramatically cheaper than they are. We are going to use that advantage just to grab share, grab more share. After we do that, we want even more share. We want momentum. We want to get into a leadership position, and we have the opportunity to do that. We are a pretty profitable company prior to being in the infrastructure business. We are just barely in the infrastructure business. We are a very profitable company already.
We think this is going to be a hugely profitable business for us, even at these prices. Our performance advantages translated to price advantages, which are significant, we think we can make huge margins. Now, the margins in the infrastructure. I mentioned the margins and the incremental margins in the SaaS business is 90%. The margin in the infrastructure business, it depends exactly on what you are buying, but it is 50% even at these prices, which is the company margin we are shooting for is 50% margins. This is not going to dilute our margins. We will take it. We think we can get a lot of people to move to our cloud with those prices. Yes, sir.
Hi, Larry. It's John DiFucci from Jefferies. The emphasis today has been Gen 2 and the cloud, and you say yourself, you're not sure when the Oracle Autonomous Database is going to take off. You have some confidence it's next year.
I can't pick the quarter in 2019.
Okay.
I think you'll get data points in 2019. You'll see a little bit of inflection, then it will go up. There'll be some surprises in 2019, for sure, which are infrastructure revenue surprises.
Okay.
You'll see the beginnings of this.
Okay. As you know, your customers, large enterprises, one of the things you can't tell is sometimes they look at this too, and you believe it, and if they believe it, sometimes it still takes time for them to move. I just wonder, I've been waiting for them to really start to adopt in mass some options that haven't been around that long, like in-memory and multi-tenancy. I don't know if we've even seen that in the numbers yet. Am I wrong to think that we will see that even before? Because that'll take time for them to move to the cloud.
Well, I think-
Different nature
No, I think in-memory multi-tenancy, which are very interesting and important features, is very different. I think people have really been focused on coming up with a cloud strategy. What are they going to do with the cloud? How are they going to do this? I think there is a sense of urgency, at least to come up with a cloud strategy. I think you're going to see people experimenting, lifting Oracle workloads and trying this workload and trying that workload and starting to move. With Gen 2 of the Cloud and Autonomous Database, we should be able to show enormous savings. Again, you'll see the beginning of it in 2019. It's very different. You can get people to talk about the cloud. It's very hard to get people to even talk about in-memory databases.
We have a faster in-memory database than HANA, which SAP is very proud of. Our in-memory technology is way better than theirs. It's not surprising. We're in the database business, they're not. I think so many of our customers are focused on developing a cloud strategy. They have not been entirely excited by what they saw from us in Gen 1, what they see in Amazon. To quote a senior engineer at Oracle, "The hard stuff hasn't moved to the cloud yet." I think these Oracle databases, some of them, these are fairly complicated applications and fairly hard to move to a Gen 1 Cloud. In a Gen 2 Cloud, it suddenly becomes a much more straightforward process to lift and shift this stuff. I think you'll see real movement, our customers, you'll see it in two forms.
You'll see the Autonomous Database at customer become very popular amongst some of our largest and most conservative banking and telecommunications customers. You'll see that, you'll see the bulk of our still high-end customers picking the public cloud option and moving to Autonomous Database. Now we have dedicated Exadata. You can get dedicated Exadata hardware, dedicated bare metal hardware. We build a barrier. We basically build you your own private network. You talk, there's barriers around the customer zones. We can set up a customer with their own Exadata machines, their own compute machines. They're not sharing with any other customer. We can provide that level of isolation to give our customers comfort. We think a lot of people are going to be going that way in our public cloud. We're optimistic.
We think now that we got Gen 2 of the Cloud working, now that we have Autonomous Database for both transaction processing and data warehouse working, that we're going to see this migration begin, and the scale of the migration by the time it's finished is actually hard to imagine. Thank you very much.
Thank you. Thank you all for attending today. We've got some slides that we'll be uploading shortly, please keep a lookout. You should have on your chairs the links to do that, we'll also get that out publicly as well. Thank you very much for attending today. Bye-bye.