You. In a moment, we'll discuss the results for our Q4 2016 fiscal year, which ended on October 30. Joining me are Gary Dickerson, our President and CEO and Bob Halliday, our Chief Financial Officer. Before we begin, let me remind you that today's call contains forward looking statements, including Applied's current view of its industries, performance, products, share positions, profitability and business outlook. These statements are subject to risks and uncertainties that could cause actual results to differ materially from those expressed or implied by such statements and are not guarantees of future performance.
Information concerning these risks and uncertainties is contained in Applied's most recent Form 10 Q and 8 ks filings with the SEC. All forward looking statements are based on management's estimates, projections and assumptions as of November 17, 2016, and Applied assumes no obligation to update them. Today's call also includes non GAAP adjusted financial measures. Reconciliations to GAAP measures are contained in today's earnings press release and in our reconciliation slides, which are available on the Investors page of our website at appliedmaterials.com. And now, I'd like to turn the call over to Gary Dickerson.
Thanks, Mike. I'm delighted to report that Applied Materials delivered record revenue and earnings in our 4th quarter, capping off an outstanding year. In fiscal 2016, we grew orders, revenue and earnings to the highest levels in the company's history and made significant progress towards our longer term strategic and financial goals. Some of our major accomplishments this year include our highest semiconductor orders and revenues since the year 2000 and record performance in display and service where both groups exceeded all time highs for orders and revenue. I want to thank all of our employees for their passion to create value for our customers and for making 2016 a great year for Applied.
Looking ahead, I see tremendous capabilities, momentum and opportunities across the company. This gives me increased confidence that we will again drive sustainable growth in 2017 beyond. In today's call, I'll start by explaining some of the key factors contributing to our record performance and describe our strategy for long term sustainable growth. I'll then provide our market view and describe why Applied has an increasingly positive outlook. I'll conclude with brief updates on our major businesses.
After that, Bob will provide additional details about our results as well as his perspective on our future outlook. At Applied, our strategy of inflection focused innovation leadership is delivering profitable growth. In both semiconductor and display, large multiyear inflections are enabled by materials innovation. Applied has by far the broadest and deepest capabilities in materials engineering. It's our ability to combine these competencies, technologies and products that really sets us apart and allows us to sustainably grow faster than the markets we serve.
To fuel growth, we focused our organization investments to deliver highly differentiated solutions that enable customers to build new devices and structures that were never possible before. At the same time, we have implemented a new operating system for the company that's driving repeatable success and increasing our product hit rate. A key element of this is our product development engine. We've already trained more than 5,000 engineers how to use these methods and as a result, we now see inflection sooner, develop solutions faster and generate higher residual value from our large installed base of tools. Over the past few quarters, I talked about 5 major market drivers that are contributing to our record performance today and will fuel growth for years to come.
These drivers are leading edge foundry and logic, 3 d NAND, patterning, advanced display and China. At our analyst event in September, we gave comprehensive updates on these 5 drivers. So today, I'll only cover 2 where we have new information to share. Let me start with patterning. At the analyst meeting, we said we expect our patterning opportunity to expand to around $3,000,000,000 by 2019, growing 2.3 times since 2012.
In recent weeks, new details about EUV adoption have been made public and this gives us increased confidence in these projections. Our model is consistent with others and assumes that EUV will primarily be used for cuts and vias in foundry and logic applications. That's about 20% of the total patterning market. For the other 80%, we see customers expanding multi patterning solutions and this significantly grows our addressable market. Pulling all this together, we believe that in the most aggressive case for EUV adoption, our 2019 patterning opportunity will still grow to around $3,000,000,000 as previously forecasted.
In scenarios where EUV adoption rates are slower than the most optimistic case, our patterning growth could be significantly higher. Looking beyond 2019, we see cuts and vias remaining the primary application for EUV and our patterning opportunity continuing to expand. Patterning is an area where Applied is making significant investments and has room to grow. We've already gained around 15 points of market share since 2012. We see strong customer pull for new materials enabled patterning processes where we have innovative new technologies and believe we can gain another 15 points of share by 2019.
My second update relates to display. We announced that our equipment has been selected for the world's first Gen 10.5 LCD TV Fab. The push to Gen 10.5 is driven Gen10.5 is driven by rapid growth in the market for large format TVs, 60 inches and above. A Gen10.5 substrate is about 80% bigger than the current Gen.85 standard allowing customers to optimize output for larger screens. In addition, demand for our OLED manufacturing equipment is broadening.
With this strength in both TV and mobile, increasing our estimate for 2017 display capital spending. We now believe customers will invest around $2,000,000,000 more than 2016. In the past few weeks, we've also increased our forecast for wafer fab equipment. We now expect 2016 wafer fab equipment spending will be at least 5% higher than in 2015, driven by additional foundry capacity and incremental investments in 3 d NAND. At the same time, our view of 2017 is strengthening.
We believe spending will be higher this year with logic and foundry investment remaining at robust levels and growth in memory spending. Looking further ahead, I'm very excited about emerging trends in virtual and augmented reality, big data and artificial intelligence and smart vehicles. These new drivers for semiconductor and display span consumer, enterprise and industrial applications and layer on top of existing demand for mobility, PCs and other consumer electronics. Our discussions with leading companies in these areas make it clear that these new applications create huge demand for memory and require major advances in silicon technology. These trends add to my view that demand for capital equipment is becoming less cyclical and normalized annual spending is increasing over time.
I will now talk about the progress we're making in each of our major businesses. In semiconductor, we are seeing robust demand across the board and as we said in September, we expect 2016 to be a strong share gain year for Applied. Growth in our leadership businesses is being driven by foundry and logic inflections and also by memory as epitaxy, implants and rapid thermal processing are adopted in 3 d NAND. In particular, we see strong growth in CMP where we've increased revenues around 60% since 2012 and metal CBD where we expect double digit share gains this year. Our leadership businesses are in a really great position to grow.
In 2016, we converted well over 90% of our development positions to volume production wins. We're also making significant market share gains in etch and CVD. Fiscal 2016 was our 3rd consecutive year of growth in CVD and 4th consecutive year of growth in etch, where revenues reached a 9 year high. Overall, our combined etch and CVD revenues exceeded $2,700,000,000 for the year. I'm very excited by our product pipeline in etch and ALD.
We are seeing rapid adoption of our innovative new solutions, including Selective Etch and Olympia ALD that together generated more than $230,000,000 of revenue this year. I'm also pleased by the progress we're making in advanced packaging. We have some great new products and based on the positions we are winning, we expect to double our packaging revenues over the next 12 months. In inspection and process control, we also delivered our highest ever orders and revenue this year. In about 20 points of e beam inspection share, around 7 points of share in e beam overall and secure the number one position in this and secure the number one position in this market.
E beam is the fastest growing segment in inspection and we are winning new applications that support sustainable share gains for years to come. In service, our strategy is to deliver more value to customers with our advanced service products. Our service teams are more tightly aligned with our product groups than ever before. Together, they are reducing ramp times, improving device performance and yield and optimizing and yield and optimizing output and operating costs for customers. In the Q4, we set new records for both orders and revenue and when we look at year on year comparisons, we've now grown our service business every single quarter for the past 3 years.
Display is also setting records with 2 large inflections driving growth. The first is large format TVs driving investment in new Gen 10 capacity. Large format TV units are expected to grow at 15% to 20% annually over the next 3 years compared to single digit growth rates for TVs overall. The second is OLED, where investment is increasing as multiple customers start to ramp this technology and battle for leadership in next generation mobile screens. At our analyst meeting, we said that our market opportunity in display is expanding significantly, up to 10 times in the case of OLED.
We are focused on building out our product portfolio to deliver the solutions our customers need to transition to new display products over the next few years. We have great traction with our new thin film encapsulation and e beam review products and have more significant new products that will be announced in 2017. In summary, I strongly believe this is Applied's time. We set new performance records in 20 16 and we're seeing the impact of the investments we've made in our organization and product pipeline over the past several years. As we look ahead to 2017 beyond, we see strength in our markets as large multi year inflections continue to evolve and new emerging demand drivers layer on top of mobility and computing.
Across the company, we are focused on extending our innovation leadership. Applied solutions are enabling customers to build new devices and structures that were never possible before. This puts us in a unique position to drive
Bob? Thanks, Gary. In September, we raised the ceiling on our expectations for our markets and our company. Since then, we've increased our 2016 WFE forecast to a range of $33,500,000,000 to $34,000,000,000 Our internal projection for 2017 WFE is $1,000,000,000 higher than that and it assumes only modest DRAM investment, China spending and NAND penetration of hard drives. So our early view on 2018 and the longer term horizon positive as well.
So while our Analyst Day was just 8 weeks ago, I already feel more confident in our ability to hit our 2019 financial model. After the Analyst Day, the slide that sparked the most interest was the one comparing average WFE spending and the standard deviation around the average for 2 periods, 2000 through 2,009 and 2010 through 20 16. What really caught everyone's attention was that one standard deviation fell from $8,000,000,000 in the earlier period to $2,700,000,000 The WFE industry has evolved since 2010 and that deserves some discussion. I'll give you the major factors that lead me to believe it's a fundamentally more stable and attractive business. First, the semiconductor industry we serve has become larger, more diversified and less volatile.
Specifically, semi content was led by PCs, which have multiyear demand cycles timed to enterprise upgrades. Now we've added mobility, which has annual consumer replacement cycles and generates more semi revenue than PCs. And today, we're laying on additional semi demand drivers including big data, IoT, cloud infrastructure, artificial data, IoT, cloud infrastructure, artificial intelligence, virtual reality and self driving cars. 2nd, the cycles of the past were mainly driven by PC DRAM, which was the largest, most competitive and most generic technology. Today, DRAM is near the bottom in spending.
There are fewer producers the designs have been tailored to servers and mobile devices, including multi chip stacks. It's no longer one size fits all. 3rd, NAND spending has surpassed DRAM and NAND continues to unlock growth potential as the gigabyte cost approaches hard drives. 4th, foundry capacity additions tend to be contract based and demand driven rather than supply driven and foundry has gone from being the smallest category to the biggest today. And 5th, our customers now add capacity on a more flexible basis, building out lines and line extensions instead of entire mega fabs.
Not only has the industry changed, but applied has changed in ways that give us larger and more diverse revenues and profits. Over the past few years, we have achieved relatively balanced market share across all of the semiconductor device types. Most of our businesses are now benefiting from 3 d NAND and our success in NAND has been a springboard to our new patenting wins in DRAM. At the same time, we've greatly expanded our service business whose revenues are up by over 30% in just the past 3 years. By 2019, this relatively stable business should deliver almost a quarter of our revenue and an even greater percentage of our operating And Applied is uniquely positioned as the equipment leader in display.
We're on track to triple our overall opportunity with new display products and we're accelerating our display pipeline by leveraging our semiconductor IP including e beam, identifying new technology through Applied Ventures and securing R and D co funding. In both semi and display, we can point to specific inflections and winning products that will drive our growth through the model period and beyond. And in all of our markets, I believe we're demonstrating systemic improvements in our ability to outperform. The better allocation spending, more efficient execution and product development engine are now culturally ingrained. What's the significance of this greater stability?
I'll mention just three things. First, our order variations now reflect seasonality more than cyclicality. We're focused on driving year over year revenue and earnings growth and we plan to stop reporting quarterly 16. To help you feel comfortable with this change, I'm telling you now that I expect our Q1 fiscal 2017 orders to be meaningfully higher in both semi and display. I also plan to detail our Q1 orders on the February call, so you can have a complete set of data for your calendar 2016 models.
2nd, this stability gives us a unique opportunity to get closer to our customers and here's why. The technology roadmaps are getting harder. So customers increasingly want to work with the most capable suppliers and we are the broadest and most capable. Today, our customers are asking us to aggressively. This costs money, but the more stable business environment gives us confidence that we can invest more in our customers' roadmaps and generate more attractive risk adjusted returns.
Number 3, the stability helps us deliver higher more predictable profits and increased cash flow. It's great for our financial stability. 8 weeks ago, we showed you our 2019 financial model. Some of you realize that we plan to generate substantial free cash flow well beyond what's needed to achieve our share count objective. We regularly evaluate the best means of returning excess cash to shareholders and we know some shareholders value dividend growth as much as buybacks.
We're evaluating our options for this additional liquidity and we're monitoring the tax policy environment for any changes that could influence our thinking about the mix. Now I'll shift gears and comment on our performance during Q4. I'll focus on our revenues and profits as compared to the same period last year. We grew company revenue by 39% year over year and more than doubled our non GAAP earnings per share. We increased non GAAP gross margin by 1.5 points year over year.
And our spending discipline helped us to increase non GAAP operating profit by 82% year over year. The non GAAP tax rate was a little lower than expected due to a favorable geographic mix. At the segment level, we grew semiconductor systems revenue by 42% year over year and segment non GAAP operating margin by 9.1 points. We increased service revenues by 13% year over year and non GAAP operating profit by 1.9 points. And we grew display revenue by 92% year over year and non GAAP operating profit by 10.9 points.
Moving to the balance sheet. We delivered strong operating cash flow of $797,000,000 in Q4 or 24 percent of revenue. We grew total cash and investments to $4,680,000,000 after returning $279,000,000 to shareholders through buybacks and dividends. The year, we distributed 106 percent of free cash flow to shareholders. At year end, about 14% of cash and investments were onshore.
The onshore balance has since increased to roughly 45%, reflecting the effects of intra company transactions. Now I'll provide our guidance for the Q1 of 2017. We expect our overall revenue to be close to the record levels set in Q4. Our expectation represents year over year growth of approximately 45%, plus or minus 3 points. Within this outlook, Semiconductor Systems revenue should be near the 15 year high achieved in Q4.
This forecast represents year over year growth of about 55%, plus or minus 3 points. Services revenue is seasonally lower in Q1, but we expect it to be up by about 10% year over year, plus or minus 2 points. And display revenue should be up by about 65% year over year, plus or minus 10 points. Our non GAAP earnings per share should be in the range of $0.62 to $0.70 the midpoint of which would be up by 154% year over year. As a reminder, Q1 of 2016 was a 14 week quarter.
Since it's the Q1 of a new year, I help you with your modeling questions. In Q1, non GAAP gross margin is likely to be up by about 1 point from 43.7% in Q4 reflecting very positive mix. For the full year, we expect gross margin to reach approximately 44%, which would bring us to within 60 basis points of the midpoint of our 2019 model. Non GAAP OpEx is likely to be $625,000,000 in Q1, plus or minus $10,000,000 Our quarterly run rate in the balance of the year may be approximately $635,000,000 which remains below the model. Our non GAAP tax rate is likely to be around 11%, which is at the model and we'll continue to buy back shares to achieve the share count target in the model.
In summary, we delivered new levels of revenue and profitability in 2016 and demonstrated the benefits of the strategy we've been working to implement over the past several years. But this is just the beginning. The industry is bigger and more attractive. Our opportunity set is larger. Our customer relationships stronger performance is sustainable and we're Our stronger performance is sustainable and we're confident in our ability to deliver increasingly predictable cash returns to our shareholders.
Now Mike, let's start the
Q and A. Thanks, Bob. To help us reach as many people as we can, please ask just one question at this time. If you have an additional question later, please poll the operator and we'll do our best to answer it later in the call. Operator, let's please begin.
Our first question for today comes from the line of C. J. Muse from Evercore.
J. Muse:] Yes, good afternoon. Thank you for taking my question. I guess, if you look back here or at least with the implied guide for calendar 2016 or for Q4, you're growing roughly 24 plus percent. And if you think about mix next year, I'm assuming growing display, growing NAND.
How should we think about your relative outperformance to the rest of the industry? Sure. In our fiscal year, in our calendar year next year, we think in semi, it's going to be a healthy year in general for the environment. We feel good about the environment. We're going to gain share.
We think we don't want to give the expect numbers right now, but we're going to gain WFE share next year in a number of places. And then display, we see display revenues up significant amount next year. And then we see the service business continue to grow. So we'll outperform the markets we're in. But additionally, we think the markets are pretty good next year.
Our outlook is WFE is up. Our outlook is that spending on cap on display equipment is up and services will be good.
Thank you. And our next question comes from the line of Toshiya Hari from Goldman Sachs.
Great. Thanks for taking my question. Bob, you talked about 2017 WFE being up about $1,000,000,000 based on modest expectations for DRAM, China and NAND. Can you dig a little bit deeper into those 3 items and give a little color as to what you're assuming in the model?
Sure. Part of what we were answering there was that we think there's wind at our back into 2018 too. So if you look at 2017, we think NAND is up next year. We think foundry is strong and we think logic is pretty good. DRAM up some next year, but what we don't see is high levels of absolute DRAM spending because this year is not too big.
What we're modeling, and we might be a little conservative, is not a big change in China spending. We think that hits more in 2018 if you just look
at buildings and shelves and stuff like that. We're
China ramps a fair amount more in 2018. We think China ramps a fair amount more in 2018. We think DRAM probably has upside in 2018, not to mention some of these altered form memories we're starting to see early now. And thirdly, NAND, by the end of next year, is probably going to get to about 700,000 wafer studs. So it's about $1,400,000 And we think all of that eventually get virtually all of that gets converted.
So we think there's still strong NAND spending in 2018.
And our next question comes from the line of Harlan Sur from JPMorgan.
Good afternoon and good job on the quarterly execution. One of the key takeaways from the analyst meeting was the order trajectory, right? So specifically, book to bill as a leading indicator of the revenue and the earnings power momentum. Past 4 years, you guys have had book to bill greater than 1 and obviously, clearly an indicator of the growth of the business. As you look at your customer and program pipeline for fiscal 2017, combined with your view on growth in flat panel and WFE industry spending growth, is 2017 likely to be the 5th consecutive year of book to bill greater than 1 for the team?
Well, what we said earlier in the call was that, we see our bookings continuing strong, but the volatility bookings quarter to quarter in particular, we're going to stop reporting bookings after Q1. We're going to give it to you in Q1. I mean Q1, we see up meaningfully. And what we mean by meaningfully is up more than strong double digits, okay, from Q4. So people have been worried that we peaked in Q3 bookings down some in Q4.
We see a strong Q1s and we see a strong year overall. So we think the momentum is still with us and our revenues are going to be up and our share is going to be up next year. In terms of quarterly variability in orders, I'm not sure. In terms of book to bill next year, we don't have clarity on the second half here. It's too early for us to tell.
But we see a strong WFE year, share up, strong spending on display. Frankly, it's more strong in both those than we saw 8 weeks ago. For instance, the large TVs are doing well, the Gen 10 stuff, the profitability at the display guys is good. So we see displays pretty darn strong actually. We see you're up to 9 or 10 people trying to make OLED displays now for mobile.
And then within WFE, we're getting a lot of pull from customers to increase production and shipments. So we think overall it's next year and we think we're going to have very healthy bookings. Whether it's where it is relative to 1, I don't know right now.
Thanks, Harlan.
Thank you. And our next question comes from the line of Timothy Arcuri from Cowen and Company.
Thank you very much. Bob, I'm sort of curious on China, how much you're putting in? I know that you think China is sort of generally more of a 2018 thing than a 2017 thing. But I'm wondering if you can give us some numbers in terms of what is embedded in your WFE forecast next year for China versus what it was this year? Thanks.
Well, I'll tell you what's in I'll give you a sense of what's in the numbers and I'll give you my qualitative assessment of the numbers a little bit. We think it's down a little bit next year, but the numbers stay it's down a little bit next year, but we think it's flat. So I think I'm probably conservative on that. So what's buried in my numbers is down a little bit, but we think there's probably upside to that, to be honest with you. Gary is shaking his head that he definitely doesn't think it's down next year.
Yes. I would say that applied, we're strong in all regions, including China. And as we've talked about before, it's one of the strongest regions in semi and display. Right now, it looks like 2017 is going to be roughly the same as 2016, but it's up a pretty significant amount from where we were a couple of years ago. 2018, we think could be meaningfully up above the 2017 forecast.
So those are my thoughts.
Thanks, Jim.
Thank you. And our next question comes from the line of Atif Malik from Citigroup.
Thanks for taking my question and great job on the quarter. Question for Gary. Gary, thanks for the update on the patterning with the base in the bulk case on EUV. And then on the display side, can you just give us a little bit of an update or preview on what some of the new products that you're working on, on the OLED side plan to achieve? Are these products meant to remove the supply chain bottlenecks?
Or are they going to target some of the cost of ownership or other technology challenges?
Sure. Yes, thanks for the question. Our strategy overall is inflection focused innovation. And what we've talked about in display is an opportunity to triple our served market over the next few years. The team at display is really, really an outstanding team, driving innovative new products, targeting major customer inflections.
And you can already see growth in display with thin film encapsulation and e beam, 2 products that we've introduced in the last couple of years that are really generating meaningful revenue for the display business. And so we're really continuing the same playbook. We're focused on inflections that are meaningful for our customers. We have very strong pull and investment from customers in these new products and high confidence that these new businesses will add meaningful revenue for our display business and increased confidence that we're going to hit or exceed the model that we showed for 2019. We'll disclose more about the specific products in 2017 next year.
And then I think Autofax asked 2 questions even though I asked for 1, and I wonder if you could comment on the EUV.
Sure. So we talked about patterning at our investor meeting in New York, and we see patterning as a great opportunity. We forecasted about a $3,000,000,000 TAM by 2019. And when we look at EUV adoption, we agree with others that the view of EUV is going to be mostly focused on v as in cuts in logic. So when you look at the leading customer roadmaps, vias and cuts in logics are is about 20% of the total patterning TAM, 80% of the TAM will continue to add multi patterning.
And that gives us increased confidence in this $3,000,000,000 estimate that we had at our investor meeting. Basically, if you take the most aggressive EUV assumption for adoption, something like 10 layers adopted at 5 nanometer logic for vias and cuts, you come up to the $3,000,000,000 estimate. Of course, if it's not in this most aggressive assumption, then the TAM could be higher. And Applied is really in a great position to gain share. We talked about a $1,000,000,000 number in our model for 2019 for Applied, And we have great confidence in that number.
We already gained 15 points of share in Patterning since 2012. We're forecasting at least 15% more between now and 2019. Really great products in etch, in selective removal, CVD, ALD, CMP, really, really, really great new products and we have line of sight to the 2019 model with many application wins and increasing customer pull for these innovative new products. Thanks.
Thank you. And our next question comes from the line of Stephen Chen from UBS.
Thanks. Hi, Gary and Bob. So I have a follow-up question on your 2017 view on foundry logic WFE. Does your 2017 view assume there'll be much foundrylogic customer reuse of 10 nanometer equipment for the 7 nanometer node or does it assume limited equipment reuse by foundrylogic? Thanks.
Sure. Let me we've had a couple of questions on peak questions and reuse. Let me give you some context on how we look at the business, then I'll answer your specific question, Stephen. We actually really believe three things. 1, that the industry is growing more and more attractive in which we compete.
Total spending is trending up and volatility is trending down in both WFE and in capital spending for displays. Secondly, our plant, I think, is trending up in terms of the breadth of our product offerings and reduced volatilities. We have the similar share across NAND, foundry, DRAM and logic. We're gaining share. We have a deep product pipeline.
So I think the second thing is the trend is the industry is looking good. The trend is applied is becoming more and more attractive and less volatile. And third, those things combined produce very predictable cash flow, which we can talk about later on the call, gives us opportunity to return cash to investors. In terms of risk that there's some replacement, reuse in 10 and 7 in next year, I think it's pretty moderate, frankly. I think 10 is not going to be a big node.
I think people are going to rush to go to 7, but 10 megabytes devices are shipping next year, 7 the year after. So the ability to migrate equipment is somewhat limited. The second thing is the comparison everybody always gives us the 20 to 16fourteen migration. That was distinct in a couple of ways. 1, you had the biggest consumer of computer chips in the world switch from one vendor to another in a node.
And secondly, you had gone from the last planar FinFET device to the first last planar to first FinFET, which really made the last device unattractive. Is there going to be reuse some type of equipment for 10 or 7 years? I don't think it's next year, and I think it's going to be much smoother than it was in 2015.
Thanks, Stephen.
Thank you. And our next question comes from the line of Krish Sankar from Bank of America.
Yes. Hi. Thanks for taking my question. I'd like a 2 part question, Bob. One is you mentioned the business is getting more seasonal versus cyclical.
When you look at the order trend, you had a dip in orders for semis in your October quarter. Is that the seasonality you're talking about on a go forward basis and along the same path when you look into calendar 2017? Can you help us understand, dig a little more deeper on how the WFE profile looks like first half versus second half for NAND, DRAM, foundry and logic?
Yes, I do believe that as we said on the call recorded the previous part of the call, I do think it's more seasonal than cyclical. We have a lot of data that shows it. So for instance, even if you look, as we said, the script we got asked the most questions about at the earnings call at the Analyst Day was this average WFE spending since 2010 is up to like $31,700,000,000 I think it is with the volatility about $2,800,000 versus 2,000, 2009, it was like $25,500,000,000 of the volatility of $8,000,000,000 I'll give you more details on that. If you look at memory as a sum, because virtually every customer makes DRAM and NAND, the average before 2010, one standard deviation, one Sigma was $6,000,000,000 on an average spend of $10,700,000,000 Since 2010, it's $3,000,000,000 standard deviation on an average spend of $12,000,000,000 when you put DRAM and NAND together. The second thing is virtually every year since 2012, I think it is, it's up pretty much $1,000,000,000 to $2,000,000,000 every year total NAND spending.
Now if you get lost in the weeds between DRAM and NAND, it's all different. But the companies manage their budgets that way. They say we're going to spend on DRAM this year and NAND that ramp this fab this year, the other fab the next year. And then on foundry, foundry spending before 2010, it was about $2,000,000,000 on an average the standard deviation was $2,000,000,000 since that's $1,500,000,000 on a spend of $12,000,000,000 and the spend pre-twenty 10 averaged 4.5 percent. So the magnitude of the extended divisions, the volatility is going down, okay?
It really is. It's seasonal, not cyclical. In terms of your specific question about, I guess, it was about next year, We're not exactly sure. I'll give you some data for us. We think our number one, Q1 bookings were up.
We think on the year WFE is up. We think our share is up. I was just looking at the data. We think year over year, it's a chance that every quarter year over year is up. So I think it's less.
In terms of the seasonality you specifically asked on foundry, historically spending on WFE and foundry, Foundry is the one that tends to be a little more seasonal because it's Christmas in Chinese New Year. They historically have taken equipment more March through August, like end of February through August. And then they might have another spurt if they need more later in the year. This year was a little different because one of our big foundry customers is trying to more level load themselves and get a head start on 10 nanometer. So we had a little bit different in terms of spending because remember, 10 nanometer is not shipping until next year.
And then the other thing that made this year a little different is there's just sustained demand for 3 d NAND. So we had some at the beginning of the year, we have a pull at the end of the year. So the seasonality, which is kind of more than cyclicality, is being mapped to a technology transition, which is very helpful too. And again, that's not so much cyclical. It's more of a technology trend that's going to go on for years.
Thanks, Mahesh.
Thank you. And our next question comes from the line of Faran Ahmad from Credit Suisse.
Thanks for taking my question. My first question is regarding the Jan quarter orders outlook. On the prior call, you had kind of indicated that orders would be flattish in the Jan quarter from October. I just want to understand like what has improved and how do you see the order trend by different segments within SSG?
Yes. I'm trying to get off my orders bad habit, but I'll try and help a little bit. What we said last quarter to make sure we put a floor because we had spiked up a lot in every metric you could measure last few quarters, orders, revenues, EPS. We want to give people a sense, we were going to continue strong that our share gains, our revenue growth, our operating margin expansion was sustainable. So what we said last quarter was, hey, orders are not going to fall for cliff.
The number is going to start with a 3 in the next 2 quarters, Q4 and Q1. And in fact, they will start with a 3. And in fact, the 3 was a little bit lot of we thought it would be a little higher in Q4 and some of those went into Q1. So we think we're up double digits in Q1, okay? So it's probably a little bit timing.
And the magnitude is probably a little stronger because where we are getting pull is on some flash devices, stuff like that, 3 d NAND. So it's strong and it's probably stronger even than we thought a few months ago.
Yes. Thanks, Mohan.
Thank you. And our next question comes from the line of Ramesh Shah from Nomura.
Yes. Hi. Thank you. It just seems that AI machine learning and self driving car adoptions coming along a lot faster than we thought at the start of the year. And we know these applications are iterative and very compute intensive.
And I'm just I'm wondering if this could be all of this could be an incremental driver next year for your memory and logic businesses?
Don't know about next year, but I would say that we are definitely increasingly bullish about the longer term. We've talked about 5 drivers in our model between now and 2019, these multiyear waves. So we're very confident in those drivers and we talked about that in the prepared remarks. But I would say definitely increasingly bullish relative to the longer term opportunity as we're engaged with some of the leading technology companies around cognitive computing, the smart vehicles, you saw a big announcement from Samsung this week with an $8,000,000,000 acquisition. There's an incredible amount of money being spent.
And when you look at these drivers, VR, AR, smart vehicles, cognitive computing, all of these different areas, you need higher performance computing. And we had the CEO of one company that was here recently, and we were talking about logic devices, what are the size of those logic devices going to be? And basically said as big as I can build them, as big as I can fit it into the reticle field. And he was also talking about the amount of memory that's going to be needed for some of these future applications. And it's much, much bigger than I think any of us could imagine.
So I would think we are much more bullish about the longer term relative to the drivers for high performance computing and for memory. I'm also bullish about the innovation pipeline that we have within Applied Materials. I really believe that the $500,000,000 per year that we've moved in terms of innovation, we are already enabling people to design devices that they could not have dreamed of building a few years ago. And I strongly believe that the pipeline we have now, the best is still on the is still coming in that pipeline. So I'm really bullish about the market.
Bob talked about where we compete being better, more stable, upward trajectory, less volatile. Applied is in the best position we've ever been in to enable all of these great big inflections to happen.
Thanks, Robert.
Thank you. And our next question comes from the line of Joe Moore from Morgan Stanley.
Yes. Thank you. I wanted to ask about your sort of multi year commentary when you sort of said you're still optimistic even beyond 2017. Can you talk about NAND in that context? And as you go through a very high capital intensity period of planar to 3 d conversions and 3 d greenfield, you'll get to the point where you're doing sort of more layer count increases, which is less capital intensive.
Shouldn't NAND turn into a headwind at some point in the next couple of years? And how do you think about that in that multiyear context of overall WFE?
Sure, Joe. Let me this is going to be historic day. I'm going to use phraseology on this in the next minute. It's never been used by a CFO in American public company before. All right.
So hold on. So I have a couple of slides in front of me that shows the demand for solid state disk drive market growth in both enterprise and PCs. So these numbers, you can argue them a little bit, but I think they're in the ballpark. So 2015, there was about 12 petabytes of demand for solid state disk drives in enterprise. That goes up to almost 120 1,000 petabytes in 2020.
So it goes from 12,000 to 120,000, okay? And on the PC side, in 2000, what I say, 2015, it's about 20,000 going to 140 1,000 petabytes in 2020. Let me give you another way to look at those. Our estimates that the penetration across all PCs, including workstations, notebooks, in 2015 was 21.8% and it was only 15% 2014% up to 31.6 percent of those have solid state disk drives going to 77.5 percent 2020. And if you go look at the enterprise servers, this year, it is about 20%, it was 14% of all of them in 2014, and we see it going to 26% in 2020.
So again, the total petabytes, I'll use that word again, has gone up over 100 1,000 petabytes from 2016 to 2020 in each PCs and enterprise. So the demand is there. They're expanding the addressable market for NAND. So other data I have seen is that for servers alone, you might have to add 500,000 wafer starts of capacity. And your average NAND factor of 100,000 costs about $5,000,000 or $6,000,000,000 say $5,000,000,000 6.
So that's about $25,000,000,000 $30,000,000,000 And upgrades are about another $25,000,000,000 to $30,000,000,000 So you could have $50,000,000,000 to $60,000,000,000 of spending, and that's cross calibrated, over the next sort of 5 years maybe, 2020 type data, 4, 5 years. So if you take it over 4 years, that's like you're going to be $50,000,000,000 that says. So it's over $10,000,000,000 a year, okay? So we've upped this year's forecast. Next year, we're over, I think, dollars 11,000,000 and change.
And I'm not sure it doesn't stay at $10,000,000,000 or more for a few years to come because we're kind of looking at the end user demand for solid state disk drives, penetrating high disk drives by market. We're looking at capital intensity and we're getting feedback from customers. And in the short term, they're shipping everything they can build. So what's the point? I think the NAND build out goes on for a while and I think people underestimate because the other thing is they're going to convert the 2 d to 3 d because they're not going to sell the 2 d and that will drive them to sell even more 3 d and hit cost points that eat into the hard disk drives even faster.
So I think there's a bunch of ways you can get a number that could be $10,000,000,000 for a while.
Great. Thanks, Joe.
And I'd like to say that I was proud to use the word petabytes today.
Thank you. And our next question comes from the line of Patrick Ho from Stifel Nicolaus.
Thank you very much. Bob, maybe if you could just clarify again the growth that you're expecting in total display spending in 2017. Is that incrementally because of the Gen 10.5 build out that you're seeing? Or are you also anticipating a further increase in OLED spending next year?
I think we're seeing a few things. In actual spending, I think probably we're up more we have a slide on that. So I think we're up probably more in TVs versus what we thought 2 months ago. We think that there's huge demand for OLED for mobile. The question is how much how fast can they ramp it?
I mean, I think it's a question of supply more than demand. I think on TV, they can ramp the big TVs and they're going to ramp them pretty aggressively, particularly China. This Gen 10.5 is ramping. There's going to be more to come after that. So we're up to over 16.5 percent next year for display CapEx.
I don't have the numbers from 2 months ago, but the growth is in both. But I think the thing that's bigger in 2 months for us is the TV, because the TV sizes are coming bigger than we thought, the individual TV sizes. And the profitability is quite strong in the TV panel business, which is a leading indicator of spending and then these big fabs. So I think they're both very healthy, but I think the increase in spend is probably a little bit more on the TV side.
Thanks, Patrick.
Thank you. And our next question comes from the line of Edwin Mok from Needham.
Hi, thanks for taking my question.
My question is on
the DRAM side. Obviously, you sound a little more conservative on DRAM, we're seeing pricing improve and some talk about customer looking into moving to next snow. Are you
just being conservative? Or have you kind of what are
you hearing from the customer? And are you seeing signs that maybe start moving on to the next node?
Node? Yes. We think DRAM is up some next year, but we think most customers who make DRAM and NAND are setting a priority on NAND right now because they're seeing an expansion in their addressable market for NAND, number 1. Number 2, it's a competitive situation that they don't want to miss that growing market. And DRAM, their pricing is pretty good.
So I think their capital goals will go to NAND.
Thanks.
Thank you. And our next question comes from the line of Jerome Rammel from BNP Paribas.
I'd like to know what is your assumption in term of installed capacity wafer starts per month for the 10 nanometer node for next year?
Could you remind me just our line was quiet for a moment. It sounded like you were talking about 10 nanometers and was it the capacity at the end of the year or something? Sorry, we couldn't quite hear.
Yes.
Can you please just repeat? Thank you.
Yes. Can you share with us what is your assumption for
the 10 nanometer node capacity that you have at the end of, let's say, Q4, 2017 for the industry?
Sure. We the chart I'm looking at, we put I'll give you what I have easily available, okay? The spending in 2016 in foundry, 59% is we think estimating 2016 is for tenseven and about the same percentage next year because you're still seeing strength at some of the trailing edge stuff. 28 and particularly 28 nanometer is pretty strong still. In fact, we think it might be up next year.
And then in terms of the split between 10 and 7, the total capacity between 10 and 7 is around 70,000 to 80,000 wafer starts at the end of this year, but 110,000 to 130,000 at the end of next year. I think the vast majority of that's 10 nanometer.
Yes. Thanks, Jerome.
Thank you. And our next question comes from the line of Mehdi Hosseini from Susquehanna.
Yes. Thanks for taking my question. You sound really positive regarding the demand drivers, and you're going to stop providing booking numbers on a quarterly basis. So if you're so confident, why not provide an annual revenue and earning target so that we can better track this multiyear cycle that you're going through?
We never have in the industry. I'm not going to start now. I mean, we don't give guidance out a year.
But you stopped providing bookings,
so But here's what we gave you. We gave you a 2019 model that gave you a WFE assumption, revenue assumption by product, share assumptions. We told you in the Analyst Day that our share gains were almost linear from now through 2019. And we told you that our service business was going to grow almost linearly by year. And we told you that display was going to be growing up to that number too.
So our growth in revenue from 2015 2016 to 2019, we think is pretty straight line. Now giving absolute numbers a year in advance, we don't have that level of precision.
Thank you. And our next question comes from the line of Tom Diffely from D. A. Davidson.
Yes, good afternoon. When you look at the next few years and a lot of the growth coming from both the flat panel OLED as well as China, what impact does that have on your both cost couple of large markets like Violett and China, what impact does ramping revenue in those particular markets due to your cost structure and margin structure over time?
Yes, I'll give it a shot. In terms of okay, if you want to look at cost structures applied, most of our sales are related to physical products we sell and they share a common factories, which are predominantly in Singapore and the U. S. So the material cost, product cost is similar. And then Taiwan is a big source for our displays.
So material cost structures and supply chains are worldwide. So wherever we sell them doesn't matter too much for us. In terms of differentiation and mix, which sometimes drive gross margins, different device customers use different mix of our hardware and products. In terms of OLED, we look at the OLED display market for our existing products as being similar margin profiles to our TV markets and by region similar. The opportunity for us on margin is the new products we're going to be introducing in the next couple of years in display, which we think are very highly differentiated and very attractive products, which have upside for us.
In terms of margins by region, it goes predominantly to mix of products. But historically, we provide a high level of service to customers in China, and so we usually do well in terms of share there in particular.
Yes. The incremental operating profit, if we look at China, is pretty good for us. We have a very strong position, really great relationships and a really great team supporting customers, both multinationals and the domestic companies. And also the incremental profit for us in display is positive for the company overall.
Great. And then, operator, I think we have time for just one more question, please.
Certainly. Our final question for today comes from the line of Jagadish Iyer from Summit Redstone.
Yes. Thanks for taking my question. Bob, I'm just trying to reconcile on the flash segment. We had a good splitting this year, but why were your orders essentially flat between fiscal 2015 fiscal 2016? I understand it's fiscal year, but what does it mean for 2017, I mean fiscal 2017 given the spending that you're seeing for the Flash?
Thanks.
We had one particular customer in 2015 that gave us orders pretty far in advance because they want to lock in capacity deliveries from us. So I think that was the aberration. I think the trend line is up. If you look at total WFE, it's up for Flash. And if you look at our sales are up and probably they're up next year too.
So it was a little bit in 2015. One particular customer in particular want to get into their pipeline in terms of the queue for production.
Thanks, Jagadish, for your question. And Bob, would you like to summarize before we close the call?
Sure. It seems like a year ago, but we only had Analyst Day about 6 weeks ago, maybe a little more in New York. And I have to say, even in the last 8 to 9 weeks, I increasingly believe that the industries we serve are becoming more and more attractive and growing and more diverse and less volatile. Gary talked about some of the demand drivers for our customers, whether it's autonomous cars, it's virtual reality, it's AI, big data. You really feel this with driving in Silicon Valley.
You see it everywhere. We had some of the CEOs of these companies come in and talk to us and it's real. It's happening. You look at some of the highest performing stocks in the space this year, that's what's driving them. So one, the industries we serve and displays also that way where it's kind of doubled in terms of spending are becoming more and more demand drivers would make the less volatile for us.
2nd, Applied has performed better. I think we are more and more becoming more and more innovative and executing better and there's compounding benefits from that. So I think that results in us in a broad range of opportunities. Our biggest problem is just picking and choosing among good opportunities at this point. So Applied is becoming more broader, more diverse and less volatile and more predictably we're going to do better, right?
And then finally, those two things combined and potentially with tax legislation that there's a really good opportunity to return superior cash returns to investors, including which we've said I think we said earlier today that we were conservative in the financial money model we showed 8 weeks ago in terms of the cash. And we kept a little conservatism in the cash in the model because we were wanted to be opportunistic if there was tax law changes that we could return even greater returns to investors. So I think the industry is becoming more diverse and less volatile and more attractive in terms of growth, more drivers. Applied is becoming that way. And the leverage down to the cash line, especially if you get some tax legislation that's attractive, is a pretty powerful combination for Applied.
Great. Thanks, Bob. And we'd like to thank everyone for joining us this afternoon. A replay of this call will be available on our website beginning at 5 pm Pacific Time today. So thank you for your continued interest in Applied Materials.
Ladies and gentlemen, thank you for your participation in today's conference. This does conclude the program and you may now disconnect. Everyone have a great day.