Welcome to Oracle's fourth quarter 2017 earnings conference call. I'd like to turn today's call over to Ken Bond, Senior Vice President.
Thank you, Kimberly. Good afternoon, everyone, and welcome to Oracle's fourth quarter and fiscal year 2017 earnings conference call. A copy of the press release and financial tables, which includes a GAAP to non-GAAP reconciliation and other supplemental financial information can be viewed and downloaded from our investor relations website. On the call today are Chairman and Chief Technology Officer, Larry Ellison, CEO, Safra Catz, and Mark Hurd. As a reminder, today's discussion will include forward-looking statements, including predictions, expectations, estimates, or other information that might be considered forward-looking. Throughout today's discussion, we will present some important factors relating to our business, which may potentially affect these forward-looking statements. These forward-looking statements are also subject to risks and uncertainties that may cause actual results to differ materially from statements made today.
As a result, we caution you placing undue reliance on these forward-looking statements. We encourage you to review our most recent reports, including our 10-K and 10-Q, and any applicable amendments for a complete discussion of these factors and other risks that affect our future results or the market price of our stock. Finally, we're not obligating ourselves to revise our results or publicly release any revisions of these forward-looking statements in light of new information or future events. Before we take questions, we'll begin with a few prepared remarks. With that, I'd like to turn the call over to Safra.
Thanks, Ken. Good afternoon, everyone. As you can see, we had a tremendous quarter in just about every way as cloud revenue, new software license, and earnings were all much better than expected. The adoption by our customers of our products and services is at an all-time high. Before going into more detail about Q4, you can also see that we have made significant improvements in our financial reporting to align with how we are running the business now. The cloud has become our predominant growth vehicle. As you can see in our press release supplemental pages, we have given you tons of detail.
We've given you the detail in our new presentation format, and we've also provided the reporting in our old format, going back through fiscal 2016 and 2017, so you can see the numbers the new way, the old way, all including history to help your transition this quarter. I'll first walk you through the reporting changes, then go over to Q4 results using the new format and recap Q4 using the old format for comparability to my previous guidance. Then I'll sum up fiscal year 2017 results using the new additional detail before moving on to my guidance for Q1. Please be patient, as I'm going to be giving you a lot of numbers today. As I said, it's all in the press release tables that are on file, so you don't have to catch every number.
After all that, I'll turn the call over to Larry and then Mark for their comments. In making these reporting changes, we are now reporting SaaS revenue separately as our tremendous growth resulted in SaaS revenue crossing the $1 billion a quarter threshold in Q4, having grown 76% in constant currency. We've also combined Platform and Infrastructure as a Service as the synergies and cross-selling between these two businesses is very high, and this is also how we are measuring our internal organization. On page 10, you can see GAAP applications revenue, which is SaaS and on-premise application software revenues, and further down the page, GAAP platform and infrastructure revenues, which is PaaS and IaaS and on-premise revenues. For hardware, we've gone ahead and consolidated hardware products and hardware support into a single line item of hardware. We've also aligned the expense lines where appropriate.
I'm now going to go over our non-GAAP Q4 results with our new format and use constant dollar growth rates unless I state otherwise. This quarter, by the way, the effects of currency movements resulted in a 1% headwind on both total software revenue, including cloud, as well as total revenue. Currency movements had a $0.01 negative impact on non-GAAP EPS and $0.02 negative impact on GAAP EPS. All right. Clearly, we're absolutely thrilled with our Q4 results, as software and cloud revenues were four points above the high end of my guidance, and earnings per share was $0.08 above the high end of my guidance. As I mentioned earlier, Cloud SaaS revenues for the quarter were more than $1 billion for the first time and up 76% from last year.
Cloud PaaS and IaaS revenue for the quarter were $403 million, up 45% from last year. You can see the continuing revenue momentum of our cloud business in the cloud billings and deferred revenue. The gross deferred revenue balance is now over $2.4 billion, up 63% in U.S. dollars. Cloud billings grew 42% in U.S. dollars this quarter. Once again, we've put the Cloud billings number up on our website for you to see the detail this quarter, and frankly, probably don't need to keep doing it since many of the numbers you can derive As our SaaS business continues to scale and grow dramatically, the gross margin has expanded. The gross margin for SaaS in the quarter was 65%, up from 54% last Q4. We expect to see further improvement in FY 2018 and remain committed to our goal of 80% SaaS gross margins over time.
The gross margin for PaaS and IaaS was 47%, down from 54% last Q4, as we invested in our geographic build-out ahead of the bulk of the revenue recognition. When we approach scale, I expect to see improvements in gross margins there, too. Total cloud revenues in the quarter were $1.4 billion, up 66% from last year. Total on-premise software revenues were $7.5 billion, essentially unchanged from last year. While new software license revenues were $2.6 billion, down 4%, reflecting the increasing preference of customers for cloud, software updates and product support revenues were $4.9 billion, up 3%, reflecting the continued high attach and renewal rates that show the stability of our installed base of on-premise customers. Total software and cloud revenues were $8.9 billion, up 7% and four points above the high end of guidance.
In addition, the total on-premise software and cloud revenues for both our applications and platform and infrastructure businesses are growing very well now. Applications' total revenue were $2.9 billion, up 10% non-GAAP, 8% GAAP. Platform and infrastructure total revenues were $6 billion, up 6%, both GAAP and non-GAAP. In particular, our database business, in aggregate, grew 8%. Hardware revenues were $1.1 billion, down 12%, and services revenue were $894 million, up 4%. Total revenues for the quarter were $10.9 billion, up 4% from last year. Non-GAAP operating income was $5 billion, up 7% from last year, and the operating margin was 46%, which was up from 45% last year. The non-GAAP tax rate for the quarter was 20%, as we saw some one-time benefits, and EPS was $0.89. The GAAP tax rate was 14.1%, and GAAP EPS was $0.76.
Because of currency movements, non-GAAP EPS was lowered $0.01, and GAAP EPS was lowered by $0.02. Q4, using our old format against my guidance, you can see that SaaS/PaaS revenues were $1.2 billion, up 75%, and above the high end of the guidance I gave last quarter. IaaS revenues were $214 million, up 29%, and at the high end of my guidance. Moving on to the full fiscal year, software and cloud revenues totaled $30.4 billion, growing 6% in constant currency, with $24 billion or 79% of it being recurring revenue, up from 75% last year. Cloud SaaS was $3.4 billion, growing 70%. Cloud PaaS and IaaS was $1.4 billion, growing 63%. Total cloud revenues totaled $4.7 billion, growing 68%.
On-premise software declined 1% to $25.6 billion, as the 3% growth rate in software support was offset by cloud-related declines in new software license, which were down 11%. Hardware revenues were $4.2 billion, declining 10%, and services revenue were $3.4 billion, up 1%. Total company revenues for the year grew 3% to $37.9 billion, and operating income was $16.2 billion, up 3%. Our non-GAAP operating margin for the full year was 43%, up slightly from last year. Non-GAAP EPS, earnings per share, were $2.74, up 6%. Because of currency movements, non-GAAP EPS was $0.03 lower. Operating cash flow over the last four quarters was $14.1 billion, up 3%. Capital expenditures for the quarter were $525 million. Free cash flow over the last four quarters was $12.1 billion.
I expect our cloud CapEx spending to be about $1 billion next year, roughly equivalent to this year's level. We now have approximately $66.1 billion in cash and marketable securities. Net of debt, our cash position is approximately $8.2 billion. The short-term deferred revenue balance is $8.2 billion. Up 8% year-over-year and up 11% sequentially. As we've said before, we're committed to returning value to our shareholders through technical innovation, strategic acquisition, stock repurchases, prudent use of debt, and a dividend. In terms of acquisitions, we always have a disciplined approach to finding the right companies at the right time, at the right valuations, and that make both strategic and financial sense. This quarter, we repurchased 11 million shares for a total of nearly $500 million.
Over the last 12 months, we have repurchased 86 million shares for a total of $3.5 billion and paid out dividends of $2.6 billion. The board of directors again declared a quarterly dividend of $0.19 per share. To the guidance. I'm going to give you guidance for Q1, then some preliminary comments for FY18. My guidance today is on a non-GAAP basis and in constant currency, but assuming current exchange rates remain the same as they are now, currency should not have an effect on total revenues or EPS. For Q1, cloud revenues, including SaaS, PaaS, and IaaS, are expected to grow between 48%-52%. Total revenue growth is expected to range from 4%-6%. Non-GAAP EPS in constant currency is expected to be somewhere between $0.59-$0.61 up from $0.55 last Q1. This assumes a non-GAAP tax rate of 23.5%.
Of course, the tax rate could end up being different. Over the full year of FY18, I expect continued high growth in cloud revenues, with cloud revenues materially surpassing new software license revenues. Operating income growth will accelerate and the continued expansion of cloud gross margins and the return of operating margin expansion. Over the last 2 quarters, we exceeded the high end of my EPS guidance by $0.14. As we begin FY18, our compare is now that much higher. Even with this higher compare, I expect EPS growth will be double digit next year. As I mentioned earlier, we will continue to invest in capital in our cloud data centers expansions, but I do expect our overall CapEx will be lower as the CapEx for real estate should be somewhat lower. With that, I'll turn it over to Larry for his comments.
Thank you, Safra. Last fiscal year, we sold more than $2 billion in cloud annually recurring revenue. This is the 2nd year in a row that we've sold more cloud ARR than Salesforce.com. We are now well on our way to passing them and becoming number 1 in the enterprise SaaS market. Our rapid SaaS growth is the driving force behind Oracle's revenue and earnings growth in Q4. The reason we're confident that we'll pass Salesforce is because we have 3 full SaaS application suites for ERP, for HCM, and for CRM, including financials, procurement, supply chain, manufacturing, human resources, payroll, marketing, sales, and service. Salesforce, in contrast, only competes in 3 of these 9 market areas. Furthermore, Oracle is now the clear leader in cloud ERP is by far the largest application market, not CRM.
In addition to our demonstrated strength in SaaS, Oracle also competes in infrastructure as a service and platform as a service. Here, our primary competitor is Amazon AWS. During this new fiscal year, we expect both our PaaS and IaaS businesses to accelerate into hypergrowth, the same kind of growth we're seeing with SaaS. As our customers begin to migrate their millions of Oracle databases to generation two of the Oracle public cloud. The AT&T deal is just the beginning. We expect that our Oracle PaaS and IaaS businesses will grow so fast that they will be even bigger than our SaaS business. Over to you, Mark.
Hey, thanks, Larry. First, I'd like to congratulate Safra for getting through all that material. That was a thing to get through. Listen, thanks, Larry. Good quarter for us, obviously. We tracked ahead of virtually every metric that I track, whether it was cloud bookings, cloud revenue, EPS, really everything on my KPI list was at a record level. Starting with cloud bookings, $855 million. It's the best quarter we have ever had. It's up 43% over what was a very strong Q4 last year. We had a goal of $2 billion in ARR. We finished with nearly $2.1 billion. Next year, we will sell more. By the way, booking revenue growth rates I give are in CD, unless otherwise said. SaaS bookings were $486 million, PaaS and infrastructure bookings $369 million. As Safra said, cloud revenue growth at 66%. We're now at a $6 billion annualized run rate.
For total software and cloud, to Safra's point, 80%, roughly 80% of our trailing 12 months revenue is now recurring in nature. SaaS revenue, while up 76%, was an acceleration from 40% growth last year. ERP was up 156% organically. That did not include NetSuite. Overall, ERP is now over a $1.2 billion annualized run rate. Fusion HCM was up 96%, more than twice the growth rate of Workday. Our front office solution sales, marketing, and service were all up double digits organically. Our verticals were up 111% and over $140 million in quarterly revenue. Our PaaS infrastructure revenue was up 45%. Database as a Service was up 62%. Our database business, including Database as a Service and our on-premise licenses and support, were up 8%. Cloud billings grew 42%, and SaaS percent in deferred revenue up 63%. Both of those are in USD.
I am going to just name a few customers so you get an idea of some of the customer wins we had in the quarter. I'm not going to hit everything. I'm just going to hit a few representative customers. I'll start with ERP. Allianz, Ball Corporation, BNP Paribas, GE, The Kraft Heinz Company, Juniper, MetLife, Minerals Technologies, Motorola, Mouser Electronics, NCR, Netflix, Newell Rubbermaid, Orange, Pearson, Sinclair Broadcast Group, Textron, University of Maryland, Vanderbilt University, Volkswagen. In HCM, ABM Industries, AutoZone, Fannie Mae, Fujitsu, Garden Fresh Restaurants, Landmark, Mary Kay, the National Football League, NTT, Raymond James Associates, Sinclair Broadcast Group, Staples, University of Pittsburgh, The University of Texas System, Vanderbilt, and a very large investment bank in New York City who switched vendors to our HCM. Those are just a few representative customers across just a couple of pillars in the quarter.
Lastly, as I mentioned, it was an overall a very strong quarter. Bookings, billings, and revenue, all at record levels. We announced an exciting deal with AT&T, and while it provided no revenue at all in Q4, it's a very strategic win as a reference to all of our customers about the modernization of databases and the movement of them to the cloud. Looking forward, just a couple of predictions. FY17 bookings were $2.1 billion. Our pipeline is very large, and ARR will be higher in FY18. I do expect our application ecosystem of on-premise, that would include licenses and support and SaaS, will grow roughly 10% next year. With revenue now at an annualized run rate of $6 billion and a growth rate of 66%, we're clearly the fastest-growing cloud company at scale. With that, we'll take your questions.
Thank you. Kimberly, if we could please start up the Q&A.
Ladies and gentlemen, to ask a question, press star one on your telephone keypad. To withdraw, press the pound key. Our first question comes from Raimo Lenschow with Barclays Capital.
Hey. Congrats on a great quarter. Quick question around the database side of software. Thanks for the extra disclosure. That's really helpful. I can see that the database license numbers are starting to accelerate. It did accelerate in Q4. Is that already 12c, or how do I have to think about them? Thank you.
Larry, maybe Larry might want to talk about 12c a bit.
Yeah, the new versions of our database are experiencing a very rapid uptick. We have a number of key features. One is in-memory, the other is multi-tenant. People are moving on to the latest version of our database more rapidly than they traditionally do. You saw evidence of that in Q4.
Our next question.
Thank you.
Comes from Kirk Materne with Evercore.
Thanks, my congrats on a great end to the fiscal year. Mark, I was wondering if you could provide a bit more detail on the strength in the ERP business in the cloud and why that seems to be accelerating. Just as a quick follow-up, I noticed you didn't provide any customer counts this quarter by pillar, so I was wondering if you might have those numbers for us as well. Thanks.
Yeah, I was trying to save time, sure, I'm glad cover it. We had a little less than 1,600 new SaaS customers, 1,575 in the quarter. We had 1,138 expansions in the quarter. Roughly 400 of those customers who bought SaaS also bought platform as a service. Again, the attach rate between apps and PaaS continues to strengthen. ERP, we had 868 new customers. By the way, Kirk, that doesn't include NetSuite. That's just inside the traditional Oracle ERP business. We had almost 200 expansions. That's a big deal because as we start to sell, in many cases, financials, we're now being able to sell supply chain procurement and manufacturing into that base as well. Another good thing for us in the quarter is roughly two-thirds of the new ERP customers never had Oracle ERP before.
Again, this is 66%, 65% brand new logos to Oracle. HCM, we had about 325, 326 new customers and 300 expansions, and a little over 600 in marketing, sales, and service, and about 670 expansions. All in, we have now about 13,550 customers in our SaaS active base. If you do include NetSuite, that number gets to 25,000. About 75% of everything we did in the quarter was Fusion. Again, that would not include NetSuite. That's roughly speaking, how the metrics worked out in the quarter. I hope that helps.
Next question, please.
Our next question comes from Sarah Hindlian with Macquarie.
All right. Thank you. I'd like to add congratulations too. That's a really phenomenal quarter around this aggressive cloud transition. I just had a couple questions I wanted to address for the team. Safra, I want to start with you. You just reiterated the FY 2018 double-digit earnings growth, and I guess I think what's becoming more relevant is, do we think about this as a one-year phenom bouncing off of this difficult transition, or how can we really be thinking about earnings growth beyond FY 2018 into 2019 and 2020? Mark, my second question is for you. Really interested in hearing some of these larger logo wins you cited in SaaS and ERP and HCM in particular, and yet you're still seeing, and we're still continuing to hear about a lot of traction for you in the mid-market, reflected in some of your other comments.
I'm trying to wrap my head around some of these larger logos that we're hearing pick up, especially on the ERP vertical, given its sheer size. I'm hoping you can guide us as to where we are with some of those large logos in terms of go live, pipeline, et cetera. That would be really tremendously helpful.
This is absolutely not a one-year phenomena. In fact, what you should see as this goes on is we will have less drag from the transition, and the base will continue to grow. This should really accelerate. Understand that in our PaaS/IaaS business, we're not even at scale. As we really scale that up, profitability is going to increase more quickly, and revenues will be built on the base of the recurring revenue business. We've been basically turning, and we've had the drag of the move away from our historic business to some extent, and revenue recognition that's not up front, but ratable. All of a sudden, you're going to see less of a drag as that side of the business is smaller compared to our cloud business. We're obviously very optimistic.
We indicated to you all during last year, this past year, that this thing was coming, and it has clearly come in Q4. This next year and years after that are now going to be building on that base.
To your other question, yeah, we've got some very large, active HCM customers. Today, we've got Airtel, I don't know, they're probably 30,000 users. American Eagle, 40,000 users. Fairmont Raffles Hotels, 42,000 users. Siemens, 60,000 users. Wells, 275,000 users. Xerox, 38,000. Schneider Electric, over 100,000. These are big customers. Sarah, those are just HCM. Those are obviously very big-scale customers. Now, that said, in ERP, you're right in what you said. We have a lot of upper mid-market wins and a lot of new logos to Oracle who were not doing business with Oracle before they made this acquisition. That's certainly been a fantastic strength. Our cloud is made up of a lot of new logos in addition to some existing customers. Now, in terms of scale ERP customers, though, Orange, they're using financials, procurement, projects, supply chain management. GE Digital, Qualcomm, Hearst, Tesco, Caesars.
These are just off the top of my head. These are big-scaled ERP customers. Now, to Larry's point, and I'll make it a little bit about HCM and ERP, the TAM in ERP is materially bigger than the TAM in HCM. The TAM in ERP is 3 to 4 times bigger in ERP, and additionally, ERP drags HCM or can drag HCM in the transaction. These big ERP references that we have, and obviously with the number we closed in Q4 and are going live now every quarter, more to come. We see an increasing pull of HCM in these big ERP transactions, and they are true in the mid-market and in the upmarket. Hope that helps.
Our next question comes from John DiFucci with Jefferies.
Thank you. These are impressive results, and I'll let others focus on those details. Mark, I'd like, if you could, talk a little more about the AT&T deal. AT&T, as you pointed out in the press release when that came out, has been a customer of Oracle, and we all know that. I guess the question is how incremental is this deal to Oracle? Is it a price uplift because you're providing the entire infrastructure with cloud versus just the software? Does it also include new workloads? I guess, can you give us any idea about how much incremental business, assuming there is some, which I think there obviously is, this means for you versus your previous relationship with AT&T?
Okay, first, I'm not going to give you really all that many details from a financial perspective. Is your point, was AT&T a good customer before this happened? Yes. Is AT&T going to be an even better customer going forward? Yes. Is AT&T giving us more revenue than before? Yes. Did AT&T give us any incremental revenue in Q4? No. Is this deal ratable over time? Yes. Is it a long-term agreement? Yes, it's a very long-term agreement. Let me try to give you a little bit more color. First, it is a long-term deal. AT&T has over 10,000 Oracle databases. They have several hundred. You might say, "John, that's a small percent of their overall databases," but they have several hundred large databases that have 70%, 75% of all of their company's data. They badly wanted to get the benefits of cloud.
Provisioning, auto-provisioning, all the new features that come with the product, modernization, consolidation of that infrastructure. In many cases, they've got regulatory pressures on what they can put in the public cloud and what they can't. This is an example where we've talked about this before. We take our Oracle Cloud Machine, we are able now to do all of that with them on their premise and give them all the benefits of the cloud. We manage, we patch, we basically run the cloud for them, and we help them get all of that done. As great a deal as this is, this is even a bigger deal, John, in what it represents for our entire customer base. This is a bigger deal to us than the transaction.
It is the opportunity for us to now do this for hundreds and thousands and tens of thousands of customers. The opportunity for us in the quarter, and the reason we somewhat put the press release out was AT&T being so effusive about their opportunity now and what it meant for them internal to AT&T. It's just, if not more important for us in what we can do for all of our customers. Just to give you a little bit more data, not in any of our hardware numbers, is we did in excess of 100 Oracle Cloud machines in the quarter, which are, again, us bringing all of the capabilities of our public cloud to our customers in their data centers. That's some details, although not the financial details of the deal.
Frankly, John, I think it's just as important what it means to our customers going forward on the bigger picture level.
If I could just say one other thing to John. The issue is really for the folks who have a large installed basis of Oracle Database, and they've done the research of where those workloads should be, they pick Oracle. That's really what this shows. The folks who could go anywhere and do anything come to us, and that's really the important message, and we're seeing that throughout the customer base, and you're going to start to see it in the revenues.
Next question, please.
Our next question comes from Heather Bellini with Goldman Sachs.
Great. Thank you very much. This question could be for Mark or Safra. Obviously, the cloud bookings number was very strong. I was wondering if you could just share with us any color on the impact that NetSuite might have had on the growth rate for cloud bookings. Also, while I know you're not guiding to fiscal 2018 cloud SaaS PaaS growth, I'm just wondering if you could share with us, Mark, you commented about having an expectation to do better than $2 billion in new cloud ARR. I'm just wondering if you could talk a little bit about your confidence level there. Then just for the sake of clarity, since we do get this question from time to time, this $2 billion plus number that you mentioned is all new cloud business and does not include renewals. If you could just clarify that for people.
Thank you.
Well, that's great. You gave me one question with five parts. I'm going to try my best to remember all your questions and answer. First, our bookings are new and expansion bookings. They do not include renewals. There are no renewals in the ARR that we report. Okay, point one. Point two, how did NetSuite do? I think they did terrific. I think they just did terrific. I think they're very stable. We've seen some acceleration in their growth rates from when they were a public company. I think they've just done terrific. We're thrilled with the acquisition. We're thrilled with how they're performing. We're thrilled with the synergy inside Oracle. I want to make sure I'm clear, that's not rhetoric. We're really thrilled, but I have no facial expression I can give you so you can see it. Our pipeline is big, yeah.
How confident am I in more cloud bookings than this year? Extremely, put quotes around extremely, underline it. That said, I think they're going to have a terrific year in the context of a terrific pipeline. I think I covered all your points.
Our next question is from Philip Winslow with Wells Fargo.
Hey, thanks, guys, congrats on really a fabulous fiscal Q4. Thank you for the increased disclosure in terms of the on-premise breakdown of apps versus platform. When I look at your Q4 results in the context of the full year, it's basically almost impossible to argue that Oracle's not at the re-accelerating revenue growth inflection point in its transition to the cloud. Mark, when you talked about apps, obviously it's been growing in mid to high single digits, which is acceleration from last year. You talked about accelerating to 10% this coming fiscal year. My question to both, it's either you or Safra, is when you think about the platform side of the business here at Oracle, that's been growing in the low single digits.
Where does that business fit in terms of its life cycle transition to the cloud, in terms of just the trajectory of net revenue growth? I just have one follow-up to that.
Give us the follow-up first, and then we can pace ourselves through the answers.
Well, the idea would be is if the PaaS side continues to take off, and I want to put sort of in the context of Safra's comments about CapEx being lower this year. The question I get is that, hey, if this transition to the cloud is working, why is Oracle's, call it, operating model different in the cloud? Why is Oracle so efficient in terms of its, call it, infrastructure spend in the cloud? Maybe that's a good question for Larry, actually.
Yeah. Great. Okay.
All right. Well, let me just talk a little bit about PaaS and IaaS together. The old IaaS was, as you know, a slower growing business. Our Gen 2 IaaS, much faster. PaaS, much faster. As it gets sort of completely overwhelms the base of our old IaaS, you're going to see acceleration. You must see acceleration because it is inevitable as we do more and more transactions and bookings, which turn into revenues as people bring their Oracle licenses, their Oracle workloads to us. You're going to see that's a very big area. As far as margins, remember, in our PaaS business or our IaaS business, the system itself is the same. However, what you see, though, is that there's enormous value in both our automation and in our technology, whether it's the different pieces of software that are in there.
Depending on the mix, the margin can be extremely high. It's still the Oracle you know, in that we watch every nickel around here, and we all sign off on every data center expansion and all the hardware that goes in it. We really are very careful on not getting too far ahead of our revenues, but also benefiting from massive economies of scale. This past year, we did a lot of build-out for which the revenue has not been showing up yet in the financial statement, but it will be. Depending on how the different pieces roll in, we expect margins, by the time it's the end of the year, we expect even PaaS IaaS margins to be better. We do feel we can do it within our CapEx envelope.
Okay. Let me comment. What is different? Why do we not have these huge build-outs and then wait for the revenue? We have a system. We build our own hardware. We build virtually all of our own hardware. Not some of the network switches, we buy those from Arista. We don't ship computers to any of our data centers until we have pretty firm orders for capacity. We're constantly measuring how much our data centers are used. As our data centers reach, let's say, I'll pick a number, 75%-80% utilization. We get to some threshold, we start to replenish and add compute and storage capacity to the data centers. It's a just-in-time system. Those components grow and go to Frankfurt or go to London or go to Singapore or go to any of our data centers around the world.
We do it in such a way that once we ship, we know that we're going to have a very high utilization rate at our data center. It's a just-in-time shipment, and we only ship when they cross something like a 75% or 80% utilization at that data center. We don't have a lot of unused capacity. If, let's just say, if our CapEx was double what Safra forecasts next year, that's great news because then our PaaS revenues and infrastructure revenues would be much higher and make up for it with much higher profits. I don't know if that's clear. We don't make it until, in a sense, we've already sold it.
Right.
Last point, first of all, just to add to everything Larry said. Virtually most of the COGS, other than what Larry identified, that sits in the data centers is ours. All of our R&D is aligned to all of our COGS, we have both the advantage of being the major supplier to ourselves, in addition to all of the R&D that we have into our technology. We get multiple benefits in addition to what Larry described. Phil, just to be clear, I just want to make sure I'm clear on my quote. I said roughly double digits on the application ecosystem growth next year.
All right. Let me add one more thing. How can we do it for less? We build these storage systems that Amazon doesn't have and Microsoft doesn't have. Those storage systems are really a combination of flash and disk storage. We have these huge flash caches in front of the disk. From the customer standpoint, they think almost all the data is coming from flash, and it has flash response times and flash performance. We can do that, we can deliver that with hard disk costs. We can deliver that at about, I don't know, 20% the cost of our competition. Storage is a huge component of what cloud providers are selling, and we just have a better mousetrap. We have a better storage hierarchy system than they do.
Therefore, we can provide very high performance at a dramatically lower cost to us than our competitors.
Our next question comes from Kash Rangan with Bank of America.
Hi. Thank you very much. Congratulations. I'm curious if you could talk about what drove the strength in license revenue this quarter. Certainly much better than most people's expectations. Did you have any products that were phased out or were comped out that you had any sort of pull-in potentially? Secondarily, Safra, you talked about long-term SaaS gross margins approaching potentially 80%. If this AT&T deal is a harbinger of more databases transitioning to a PaaS model, what does the longer-term outlook for gross margins for the entirety of the cloud business look like? Finally, I had to sneak in one for Mark. Any thoughts on the sales force as you approach fiscal 2018? Any changes, reorg, et cetera? That's it for me. Thank you.
That has to break every rule in the question etiquette thing. I think first of all, let me go to the license quarter. There is no standout event. There is no product that got discontinued. There was no single transaction. There was no geography. I would say North America, I will say, first of all, I'm overall very proud of our sales org in general. North America was very strong again. I think they did a great job across both of our teams in North America. We were solid really across every geography. There was just no real. Asia was strong for us, Japan. I could go on. There was no deal, there was no product, there was no geography. It was just really a broad-based performance overall. Do you want to talk about the gross margins?
Yeah. Okay. SaaS gross margins will continue to increase all of this coming year. We will get closer and closer to my 80 goal by the end of the year. We probably won't hit 80, but we'll be pretty close by Q4 in SaaS margins. I think that's good news. As a result, margins for the entire cloud business will be up quite a bit by the end. I'm very upbeat, and I expect margins to improve a lot.
I forgot your sales reorg question. That's, again, I saw a couple of notes on that. It's just like we're laying off salespeople. This is all just not true. Is our sales force shrinking? No. Do we change things in our sales force? Sure. Is there any major reorg going on? No. Remember, our competitors change. We have some very simple principles how we run the sales force. We line up our sales force by product. We line them up by buyer and by competitor. Those dynamics can change. Our competitors change, we've noted, over the past several years. You'll see some shift in our specialization. That's one point. Second, we are building out, as we've told you over the past couple of years, big hubs that are doing a lot of our selling.
The byproduct of that is that our sales force is actually increasing in numbers. Our cost per salesperson is actually declining. Our overall productivity is actually inclining. Those phenomena occur simultaneously. Anything that you hear about we're reducing our sales force, again, fake news, nothing that's going on here. Second, we are increasing our sales force. We continually realign our specializations based on competitor, based on product, based on buyer, and we continue to build out and are very happy with the continued progress of our hubbing strategy. Hopefully, that answers your question about the sales force.
Our next question comes from Keith Weiss with Morgan Stanley.
Excellent. Thank you guys for taking the question and also very nice to FY 2017. Safra, as we're thinking about FY 2018, I think a big question investors are going to have is on the license revenue side, the declines that we're expecting, should we be expecting something more in line with what we saw in Q4, like that 5% range in terms of decline? More so should we be looking at FY 2017 the full year, that 10%-15% decline as our benchmark in terms of how that license revenue decline is going to progress throughout FY 2018?
Hey, Keith. I think this is quite cute. You know I didn't give guidance on new software license growth because it's become less and less relevant
I don't actually intend to give it now, but I'll give you a chance to ask another question if you've got one.
Sure. Shifting just to repurchases.
Yeah.
Perhaps the levels of repurchases that were done in Q4 have dropped down from what we've seen in FY 2016 in the first half of the year, probably having to do with where U.S. cash is versus overseas cash. Can you give us an idea of how you guys are thinking about those repurchases into play?
Sure. Absolutely. Love to. As you guys know, as I told you, we spent quite a bit of money on NetSuite, I needed to pay that back. I've got a couple quarters of still paying that back. As a general matter, we like to buy back quite a lot of our stock, significantly more than at the $500 million a quarter that we've been on since we bought NetSuite. Once I've paid all that back, assuming that I don't buy anything else big, I expect that we will do what we always do, which is send the money back to you all, because it's your money. We will buy out your partners, share repurchases should go up. Our stock price is going up also, so we may be spending more.
Lucky us, lucky all of us, the stock price is going up, which means sometimes I will get less shares. Once I've paid back NetSuite, I do expect to spend more than $500 million a quarter, assuming I don't buy anything else, anything else really significant. I don't know how many shares I'll get back, because of course, I don't know what the stock price will be at that time.
Our final question comes from Mark Moerdler with Sanford C. Bernstein.
Thank you. Congratulations on the strong quarter.
Thank you.
Given software support is continuing to grow, would suggest that the cloud is predominantly net new workloads from either new or existing customers. We would expect that at some point soon, your customers are going to start moving existing workloads to the cloud. A couple parts to the question. When do you expect Oracle Apps and Oracle Database support revenue to start moving to the cloud? Which moves first? What do you expect the impact will be? Lastly, what of the migration is baked into your confidence for FY 2018 cloud ARR?
Well, all of it's baked into our view, there's no new news here. Let me flip the metric on you. If 65%, which is not an uncommon number we've given you over the past several quarters, is coming from net new logos, then the flip of that is 35% is coming from our existing base. We have seen our existing base move. They've just, to your point, moved slower than what the new logos have. I actually think this is a good thing, because I believe that we will get all of our existing customers on-prem, roughly speaking, moving to our cloud infrastructure over time. In addition, when we do move, and I want to make sure I'm clear and say it again, because we now have a lot more data than I would have had two years ago.
When a customer who's on-prem paying us support moves to the cloud, they pay us more money. They don't pay us one to one, they don't pay us two to one, they pay us more like three to one. In some cases, more than three to one. In some cases, we also get the attach of platform to the SaaS that we get as well, and that is not included in the three to one that I'm giving you. We are actually very positive about the fact that if people move from our application support to our SaaS revenue, we get more money. To Safra's point about what's going on with our gross margins, we now produce better performance at the bottom line because we get more revenue at a very high margin rate. I expect that to continue.
As we get more references, Mark, in ERP, et cetera, you're going to see that accelerate. We want that to accelerate, that acceleration is a good thing for our customers, and it's a good thing for Oracle shareholders. The reason we made the deal that we made, a big deal we made over AT&T, is because of the implication that has on our database business. Safra couldn't have said it better. This is now a customer who is now a reference in the public domain explaining why. Now I think you will see that same migration as time goes on. We have more customers like them who are in the middle of the process of moving now. I gave you an example of how much infrastructure was bought in Q4 of Oracle Cloud Machines. That is one surrogate for what's happening.
If that occurs, we will get more revenue as well, I predict we'll be able to deliver higher levels of service, more functionality, more capability, more modernization, as we do it. I think you've begun to see it in the application business. We love what we see in the applications business so far, I think you're at the beginning of seeing it in the database business as well.
Thank you, Mark. A telephonic replay of this conference call will be available for 24 hours. Dial-in information can be found on the press release issued earlier today. Please call the investor relations department with any follow-up questions from this call, and we look forward to speaking with you. Thank you for joining us today. With that, I call back to the operator for closing.
Thank you for joining today's Oracle fourth quarter 2017 earnings conference call. We appreciate your participation. You may now disconnect.