Good day, ladies and gentlemen, welcome to the Q4 2018 Intel Corporation earnings conference call. At this time, all participants are in a listen-only mode. Later, we will conduct a question and answer session, and instructions will follow at that time. If anyone should require assistance during the conference, please press star then zero on your touch-tone telephone. As a reminder, this conference call is being recorded. I would now like to introduce your host for today's conference, Mr. Mark Henninger, Head of Investor Relations. Sir, you may begin.
Thank you, operator, welcome everyone to Intel's Q4 and full year 2018 earnings conference call. By now, you should have received a copy of our earnings release and the earnings presentation. If you've not received both documents, they're available on our investor website, intc.com. The earnings presentation is also available in the webcast window for those joining us online. I'm joined today by Bob Swan, Intel's Chief Financial Officer and Interim CEO, Venkata Renduchintala, Group President of the Technology, Systems, Architecture and Client Group, and Chief Engineering Officer, as well as Navin Shenoy, Executive Vice President and General Manager of the Data Center Group. In a moment, we'll hear brief remarks from Bob, followed by Q&A. Before we begin, let me remind everyone that today's discussion contains forward-looking statements based on the environment as we currently see it, and as such, does include risks and uncertainties.
Please refer to our press release for more information on the specific risk factors that could cause actual results to differ materially. A brief reminder that this quarter we have provided both GAAP and non-GAAP financial measures. Today, we will be speaking to the non-GAAP financial measures when describing our consolidated results. The earnings presentation and earnings release available on intc.com include the full GAAP and non-GAAP reconciliations. With that, let me hand it over to Bob.
Thanks, Mark, thanks to Navin and Murthy, who will participate in the Q&A later in the call. Before I get into the results, I'll take a minute to address what I expect is a top-of-mind question, the status of Intel's CEO search. The board continues to evaluate candidates for what I believe is the biggest and best open job on the planet. They are proceeding with a sense of urgency while also ensuring that they make the right choice for this great company. Meanwhile, Murthy, Navin, the entire management team, and 107,000 employees have come together as a team to continue driving Intel's transformation to a data-centric company. Our 2018 results demonstrate the progress we've made, I'd like to share those results with you now. 2018 marked Intel's golden anniversary. It was a truly remarkable year for a remarkable company.
Full-year revenue grew 13% and crossed the $70 billion mark for the first time, setting an all-time revenue record for the third consecutive year. Our data center, Internet of Things, programmable solutions, memory, Mobileye, and modem businesses each set all-time full-year revenue records. 2018 was also a pivotal year in Intel's transformation to become a data-centric company, pursuing an expanded greater than $300 billion market opportunity. Intel's collection of data-centric businesses grew 20% in 2018 after adjusting for McAfee. The largest of our data-centric businesses, the Data Center Group, delivered record annual revenue of $23 billion, up 21% year-over-year, on strong cloud demand and growing share with communication service provider customers. Our PC-centric business achieved 9% growth in 2018 as the PC market stabilized, and we gained share in modems. While 2018 was a record year, we expected a stronger finish.
Q4 revenue of $18.7 billion was up 9%, but short of our expectations as a result of a dramatically weakening modem demand, lower overall growth in China, cloud service providers absorbing capacity, and a weakening NAND pricing environment. While revenue fell short, we exceeded our EPS outlook by $0.06, or about 5%. We continued to deliver outstanding new products for our customers and previewed new innovations that position Intel to compete and win for years to come. And we also made significant progress in growth areas like AI, autonomous driving, and 5G. I'll take a few minutes to give some specifics before we dive into the financial results. Over the course of the quarter and culminating at CES, we highlighted breakthrough innovations that will be central to our product leadership for years to come.
We outlined our product design philosophy, which combines six pillars of innovation, process technology, architecture, memory, interconnect, security features, and software, to consistently and reliably deliver leadership products that solve our customers' most challenging problems. One example of this design philosophy in action is our unique Foveros 3D packaging technology. Foveros enables active stacking of logic chiplets for the first time in the industry's history and lets us mix and match process technologies and architectures to deliver breakthrough products. The first such product, Lakefield, is slated for production in 2019. Lakefield features a 10-nanometer hybrid CPU architecture combining a Sunny Cove CPU core, four low-power Atom CPU cores, Gen11 graphics, and more in a dime-sized product that enables the smallest PC motherboard ever possible. Silverails gives us tremendous design flexibility and paves the way for a myriad of devices and systems, combining high performance, high density, and low-power silicon process technologies.
In the data center, we began shipping the new Cascade Lake family of high-performance Xeon processors with DL Boost for accelerated AI performance, hardware-based security mitigations, and the first implementation of Optane DC persistent memory. Many of our data center OEMs and cloud customers are now offering early trials of Intel Optane DC persistent memory, which is enabling entirely new usage models and improved system performance. In client computing, we launched our new ninth gen Intel Core desktop product lineup for gaming and content creation, growth segments that demand Intel performance. We also previewed our upcoming 10-nanometer Ice Lake client CPUs, which will deliver unprecedented levels of integration, including DL Boost inference acceleration, Wi-Fi six, Thunderbolt three, and Gen11 graphics, our first integrated GPU with a full teraflop of performance.
Our 10-nanometer yields continue to improve. Ice Lake remains on track to be in volume systems on retail shelves for the 2019 holiday selling season. In Q4, we also made important progress in AI, 5G, and autonomous driving. For artificial intelligence, we saw accelerating adoption of OpenVINO, our open-source toolkit for neural network optimization, and the rapid deployment of AI-based computer vision. In addition to the strong adoption of OpenVINO by the developer community, we also launched several new products during the year, including our third-generation vision processing unit. Our partners created a catalog of AI-based vision accelerator cards with our VPU and FPGA products. While digital video was once a vertical within the IoT business, AI-based machine vision is becoming a critical horizontal capability that cuts across all IoTG verticals.
Our leadership portfolio, both hardware and software solutions, is removing the barriers to deployment and accelerating IoTG's growth. We also highlighted a new AI product on our roadmap, the Nervana Neural Network Processor for Inference, or NNPI, which is designed to accelerate inference workloads and achieve the highest performance per watt in the industry. We expect NNPI to be in production this year. 5G is another big opportunity for both our PC-centric and data-centric businesses. At CES, we unveiled the new 10-nanometer base network system on chip, code-named Snow Ridge, developed specifically for 5G wireless access and edge computing. Snow Ridge will bring Intel architecture into wireless access base stations and allow more computing functions to be distributed out at the edge of the network.
We expect to be in production on Snow Ridge in the H2 of this year, which is also when we'll deliver our first 5G modem, the Intel XMM 8160 5G. In autonomous driving and ADAS, Mobileye's effort to lead this revolution continues to build momentum with 28 new design wins and 78 vehicle model launches in 2018. In the Q4, we announced plans to commercialize mobility as a service in Israel with Volkswagen and Champion Motors, making Mobileye's breadth of products, technologies, and services unmatched in the industry. Mobileye's products now span from open ADAS and AV compute platforms to turnkey vehicle retrofits, to ultimately, mobility as a service, and they are enabled by the industry's best vision algorithms and driving policy software, the groundbreaking RSS model for AV safety, and REM real-time crowdsourced maps. At CES, we announced important progress for both RSS and REM.
ITS, China's leading industry organization for transportation standards, approved a proposal to standardize RSS for the China market. We also completed the mapping of Japan's highway system, 25,000 kilometers of roads, using data harvested from a customer fleet of vehicles outfitted with EyeQ4 in just 24 hours. This is a task that would have previously required thousands of hours of driving and scanning using specialized vehicles. This sort of breakthrough is possible only with Mobileye's combination of technology and massive market scale, and it positions Mobileye's to monetize ADAS and AV technology long before Level four and five autonomy are deployed at scale. REM will be monetized in areas beyond autonomous vehicles. We just announced a partnership with Ordnance Survey to use data collected via consumer vehicles outfitted with EyeQ4 to help utilities manage infrastructure.
Looking back at 2018, it is abundantly clear that Intel employees, the unstoppable engine driving our innovation, are more determined than at any point in our history to make Intel technology the foundation for the world's most important innovations and advances. Not only was it a record year from a financial perspective, we achieved major milestones in terms of our diversity and inclusion goals. We reached full representation in our U.S. workforce two years ahead of our plan. We also achieved gender pay equity across our global workforce. To celebrate 50 years of Intel, more than 68,000 employees volunteered approximately 1.5 million hours in the communities where we operate. I'm proud of what Intel employees achieved in 2018, and I'm equally proud on how they responded to challenges. With that, let's turn to the financial results.
The Q4 closed a record 50th anniversary year with strong data-centric and PC-centric growth. Revenue for the quarter was $18.7 billion, up 9% year-over-year. Our data-centric businesses were collectively up 9%, and our PC-centric business was up 10%. Operating margin of 35% was approximately flat with strong mix and continued spending leverage, offset by 10-nanometer cost and growth in our adjacent businesses. Strong business performance, spending leverage, and a lower tax rate resulted in non-GAAP net income of $5.9 billion, up 14% year-over-year. EPS of $1.28 was up 18% year-over-year. For the full year of 2018, we generated $14.3 billion of free cash flow, returned $16.3 billion to shareholders, including $5.5 billion in dividends. We purchased 217 million shares and increased our buyback authorization by $15 billion.
Free cash flow was $1.2 billion short of our October expectations, due largely to an increase in accounts receivable. To summarize, we had a strong quarter and fantastic year with full year revenue up 13%, or nearly $6 billion higher than our original forecast in January. Earnings per share was up 32%, and free cash flow was up 38% over last year. We're expecting another record year in 2019. As a result of our continued growth, we are raising the dividend 5%. Our leadership products continue to win share in our expanded TAM, as both our data-centric and our PC-centric businesses continued to grow in the Q4. Our data-centric businesses were up 9% for the quarter as customers choose our performance products to move, store, and process more data faster from the cloud to the edge.
Our PC-centric business was up 10% as we saw continued strength in the commercial and gaming PC segments and regained modem share. Moving to earnings, we generated solid EPS expansion in the quarter, up 18% year-over-year, and our operating income increased $580 million, with operating margin approximately flat year-over-year in the quarter. Our EPS improvement was driven by growing demand for higher performance products in the data center and client businesses, leading to higher volumes in ASPs, continued spending leverage, a lower tax rate, and lower share count as a result of buybacks. Our focus on operational efficiency continues to produce strong results, with 2018 spending as a percentage of revenue at 28.6%, down over seven points since 2015 and meeting our 30% commitment two full years ahead of our goal.
R&D is up $1.4 billion over the same period as we continue to increase investment in areas that will drive growth in our expanded TAM, such as product leadership, artificial intelligence, and autonomous driving. Over the last three years, we've grown annual revenue by more than $15 billion while adding less than $250 million in spending, resulting in a more than 25% increase in revenue per employee. Now some Q4 performance highlights by segment. The Data Center Group delivered another greater than $6 billion revenue quarter in a growing storage and networking CPU TAM that is greater than 30 million units. Revenue of $6.1 billion was up 9% year-over-year, but below our October expectations. Year-over-year growth decelerated as all three major verticals within DCG were impacted by weakness in China demand and as some CSPs moved to consume capacity put in place earlier in the year.
Platform unit volume was up 9%. ASPs were up 1%. Our Xeon ASP grew mid-single digits as customers continued to transition to Xeon Scalable in a richer mix of higher performance products. We also saw ASP expansion in our SoC products and much higher SoC volume as we continue to have success in network transformation. The higher SoC volumes resulted in more modest blended platform ASP growth. Non-CPU adjacencies were down 2%, driven by several large one-time deals in the Q4 of 2017 that did not repeat in the Q4 2018. Cloud revenue grew 24% year-over-year, decelerating from Q3 2018. Enterprise and government revenue declined 5% year-over-year on a very challenging compare versus Q4 2017 and weaker China demand.
Comm service provider revenue grew 12% year-over-year on continued MSS gains as customers choose to virtualize and transform their networks on Intel architecture. Our other data-centric businesses, IOTG, NSG, and PSG, achieved solid growth in Q4. Together, we're up 9% year-over-year, or 13% excluding the Wind River divestiture. Our Internet of Things business had revenue of $816 million, down 7%, or up 4% excluding Wind River, with operating profit of $189 million, down 27% year-over-year, due primarily to supply constraints. Mobileye revenue was $183 million, up 43% over last year as design wins and ADAS adoption continue to accelerate. Our memory business delivered revenue of $1.1 billion, up 25% year-over-year due to strong data center growth and continued Optane adoption, offset by a weaker NAND pricing environment.
The shift of our data center and client SSDs to our 64-layer 3D NAND continues in both the data center and client businesses, with volume mix greater than 75%. NSG was approximately break even for the year. As expected, Micron exercised its right to call Intel's interest in our joint venture IM Flash Technologies. This announcement does not change Intel's plans in the coming quarters, and the close of the call is at our discretion up to one year after the date of the call, with a supply agreement that extends beyond the close. We have manufacturing options available and have been shipping a broad portfolio of Intel Optane technology products for more than a year. We will continue to expand our product line and lead the industry with this exciting new technology.
PSG's revenue came in at $612 million, up 8% on strength in the data center and comms segment. We saw continued momentum in PSG's data center segment, up 50% over last year. In the advanced products category, our 28, 20, and 14 nanometer solutions grew an outstanding 70%. Operating profit was $162 million, up 4% year-over-year. Finally, the Client Computing Group delivered another outstanding quarter with revenue of $9.8 billion, up 10% year-over-year. Commercial and gaming demand continued to be strong. The notebook segment grew 8% year-over-year. The desktop segment grew 3% year-over-year. Supply remained constrained, particularly at the value end of our product range. We are working closely with our customers to align demand with available supply while we add capacity, and we expect supply-demand balance to improve by mid-year.
Client adjacencies grew 45% year-over-year, driven primarily by increased modem share gains, though modem revenue fell significantly below our expectations as a result of weaker smartphone demand. Our operating profit grew $402 million year-over-year, with operating margin up one point. While our PC volumes were down 2%, our leadership product performance and segmentation contributed to strong mix. The investments we have made in the business organically and through acquisition are delivering excellent cash flow generation. For 2018, we generated $29.4 billion in cash from operations. We invested $15.2 billion in capital expenditures and delivered $14.3 billion in free cash flow, up 38% year-on-year, and closing the gap versus EPS by four and a half points. During this period, we returned 114% of our free cash flow to our shareholders. Buybacks totaled $10.7 billion, and dividends totaled $5.5 billion.
In addition, settlements of our convertible debt reduced fully diluted shares by 40 million. To wrap up with our full year results, we ended 2018, our 50th anniversary year, with our third consecutive year of record results. Revenue of $70.8 billion, up 13% year-on-year, driven by 20% growth in our data-centric businesses and 9% growth in our PC-centric businesses. We started the year in January expecting to generate $65 billion in revenue, 30% operating margin, and $3.55 in EPS. The growth that we and the industry have seen has been remarkable. We ended the year approximately $6 billion higher in revenue, with operating margin of 35% and $4.58 in EPS. Operating income of $25 billion was up 25% on strong execution across the businesses and disciplined spending. In October, we provided a preview of our outlook for 2019.
At the time, we described a combination of tailwinds and headwinds that were balanced. The tailwinds were an expanded and growing TAM, product momentum, and business mix. The headwinds were tougher compares following an especially strong 2018, an increasingly competitive environment, and global trade. Since that time, trade and macro concerns, especially in China, have intensified. Cloud service providers shifted from building capacity to absorbing capacity, and the NAND pricing environment has further deteriorated. Those incremental headwinds are impacting our revenue expectations and slightly reducing our operating margin % forecast. The remaining factors are roughly consistent with our October assessment. Now turning to our outlook for 2019. We expect 2019 to be another record year for us, as the world's appetite for the analysis, transmission, and storage of data continues to grow.
We are forecasting revenue of approximately $71.5 billion, up 1% year-on-year, and operating margin of approximately 34%, down less than a point year-on-year. We expect a modest decrease in gross margin percentage, driven by the 10-nanometer ramp and the growth of our adjacencies. This will be partially offset by increasing OpEx leverage as we continue to make thoughtful trade-offs and invest in R&D that will accelerate our growth and profitability. We expect the full-year tax rate to be approximately 13.5% following several beneficial discrete events in 2018, and we expect EPS of $4.60. We expect gross capital expenditures of $15.5 billion, with logic spending up and memory spending down. The increase in logic CapEx reflects our effort to meet our customers' needs and avoid constraining their growth, while our investment in memory is focused on the fit-up of our independent technology development facility in New Mexico.
And finally, we expect free cash flow of $16 billion, an increase of approximately 12%. As we look to the Q1 of 2019, we are forecasting revenue of approximately $16 billion, flat year-on-year, excluding Wind River. We expect PC-centric revenue to be up low single digits on higher modem share, and data-centric revenue to be down low single digits on broad weakness in data center and continued NAND pricing pressure. We expect operating margin of 29%, down one point year-over-year, with a decline in gross margin as a result of the 10-nanometer ramp and the growth of adjacencies, partially offset by increased spending leverage. We expect EPS of $0.87, flat year-on-year. We expect 2019 to be our fourth record year in a row. We feel great about where we are and where we're going.
Five years ago, we set out to transform Intel from a PC-centric company to a data-centric company. Today, our strategy, products, and people are delivering on that ambition with strong growth, record results, and the largest TAM opportunity in the company's history. I have been inspired and humbled time and time again by our employees' commitment to this company, their colleagues, and our customers. We're just getting started. I am convinced the board will close on a new CEO in the near future, and I believe the management team, myself, and 107,000 employees will rally behind him or her to take this company to a whole new level. In the meantime, we will not be distracted by the void. With that, let me turn it over back to Mark, and we'll get to your questions.
Okay. Thank you, Bob. Moving on now to the Q&A. As is our normal practice, we would ask each participant to ask just one question. Operator, please introduce our first questioner.
Thank you. Ladies and gentlemen, if you have a question at this time, please press the star then the number one key on your touchtone telephone. If your question has been answered or you wish to remove yourself from the queue, please press the pound key. Again, that's star then one to ask a question. To prevent any background noise, may I ask that you please place your line on mute once your question has been stated. Our first question comes from John Pitzer with Credit Suisse. Your line is now open.
Good afternoon, Bob. Thanks for letting me ask the question. Congratulations on the solid results. Just relative to the data-centric guidance for the full year of up mid-single digits year-over-year in calendar year 2019, you did a good job kind of explaining what's causing the weakness in Q1 DCG and the NAND pricing. As you look throughout the year from a minus kind of low single digits to full year being mid-single digit growth. You are looking for an acceleration against what becomes harder year-over-year compares in the June and September timeframe, at least. I'm just curious if you can help us frame how you think that that re-acceleration occurs. How long of a pause are you seeing with the cloud customers?
Importantly, how are you thinking about increased competition in 2019 in the data-centric businesses, especially in the server business, as you think about mid-single digit year-over-year growth?
Thanks for your one question, John. I'll take a stab, and then Navin, share with me as well. I'm sure he has a few comments. First, I would say that as we look at just overall demand for the data center environment, the end-user demand by consumers and by enterprises, the workloads that we're seeing, the continued growth in workloads, we are as excited about the future as we've ever been. We feel pretty good about the medium and long-term trends. As you know, in 2018, our growth rate, particularly in the cloud, was up 45% through the first nine months of the year. What we saw in the Q4 was, as I'd mentioned, we saw a little bit of consumption going on.
As you know, the purchases are done in cycles, massive purchases through the first nine months, and we started to see some of those purchases consumed in the Q4. We end the year, we think, with inventory levels on the server side of the business just a little bit higher than they had been historically. As we project forward in the first six months of the year, we think that it's going to continue to be both consumption on the server side and pricing in the NSG environment to be down through the first six months. And consistent with historical patterns, we do expect the purchasing to start picking up again in the H2 of the year. That's how we see it. Medium, long-term, we feel great. Massive buying in the first nine months.
The buying slowed a bit in the Q4, and we expect that to continue through the first six months. I think the last part of your question, just competitively, then I'll kick it over to Navin. Look, we're going into 2019 with every expectation to compete to protect our share position across our entire business. We're obviously investing in the capital required to ensure we don't constrain customers' growth. We're continuing to invest in R&D. Third, we're going to invest to protect our competitive position, both on the PC side, on the data-centric side. So yeah, we expect competition to be stronger as we go through 2019, but our guidance incorporates the fact that we're going to fight to protect our position. Navin, anything else.
Maybe the only thing I'd add, John, is that from a product point of view, the dynamic to think about in 2019 is that, as Bob mentioned, we began shipping for production Cascade Lake, our next-generation Xeon. Really, that product is going to start to ramp in the middle part of the year and into the H2 of the year. The design momentum looks very strong. The product features look very compelling. The AI capability we have with DL Boost, the support for Optane persistent memory, the security hardware mitigation fixes so that the customer momentum around that product line looks very strong, but it really doesn't ramp until the middle to the H2 of the year.
As Bob said, the H1, a little bit tougher, but the H2 with product momentum, as well as what we're hearing from our customers, we expect to be better in the Data Center Group.
Thanks, John.
Thank you. Our next question comes from Stacy Rasgon with Bernstein Research. Your line is now open.
Hi, guys. Thanks for taking my question. To maybe generalize on that, if I just look through your guidance, you're basically guiding flat in Q1, and roughly flattish for the full year, which suggests that overall, you're looking for a revenue trajectory in 2019 that's very similar to the trajectory you had in 2018, which was very, very strong sequentially in many of the quarters. I guess, what is the risk, just given everything that's going on, that that may be too aggressive, especially as you had a number of drivers in 2018, both on the data center as well as on the client side, that aren't going to be repeating in 2019? How do we think about that?
Well, first, yes, 2018 was a great year, and as we mentioned earlier, it grew during the course of the year. A function, we believe, Stacy, is just the end demand for data, and we haven't seen that slow down at all. Again, on workloads, insights from our customer and the industry, workloads continue to grow. The demand for analytics, for compute, for storage, for rapid retrieval, we think only continues to grow. And we believe that we have a very good position as we go into the year, including the products that Navin referenced. So we go into the year, we think setting expectations in line with how we expect things to play out. I would say, today our outlook is a little more cautious than it was a few months ago.
We try to take into account both the macroeconomic, the geopolitical risks, the modest inventory build as we enter the year, and the competitive environment, we've taken those into account and reflected them the best we can as we go into the year. We feel pretty good about how things stack up right now, and our expectation, as we have in the past, is to deliver on the commitments we make as we kick off the year.
Thanks, Stacy.
Thank you. Our next question comes from Pierre Ferragu with New Street Research. Your line is now open.
Hi. Thank you very much for taking my question. I was surprised, Bob, on your CapEx guidance, and especially on the memory side. My understanding is that last year you spent about $3 billion there, with about half that money actually coming from your clients, not being actually Intel capital being deployed. If you had $1.5 billion of Intel capital deployed in memory last year, if I look at your guide and think logic is slightly up, memory is going to be slightly down. The actual Intel capital invested, deployed into memory this year is going to be up massively, maybe close to 2x, and that's in a year in which everybody in the value chain, everybody in the memory industry, is actually pulling back on CapEx and limiting capacity addition.
So, I'd love to understand how you see that and how you position Intel this year in memory.
First, I'd just maybe start with, as we see free cash flow for the year, we expect to be up $2 billion year-on-year, with gross capital relatively flat. I think you said this, but just to repeat, 15.5 going to 15.5. During the course of 2019, our expectations are, of that mix, it will be more logic oriented. That's really driven by a couple things. One, ensuring we have the capacity to meet the 14 nanometer demand for our customers. Secondly, as we ramp 10 nanometer in 2019 and position for 10 in 2020, we'll invest additional capital there. Then third, obviously, our expectations are to continue to invest in next node technology, in particular seven nanometer. Logic capital is going to be going up year-on-year. As we indicated, memory capital will be coming down.
We put the capacity in place in Dalian during the course of 2017 and 2018. Our expectations are, in 2019, that we have sufficient capacity for demand. However, we are going to be investing in our own capabilities or self-sufficiencies for Optane product. So we will use some capital on building out Optane capacity, but memory capital will be a bit lower. Gross capital will be a bit lower during the course of the year.
Thank you. Our next question comes from Christopher Danely with Citi Research. Your line is now open.
Hey, thanks, guys. I'm going to shift to the expense line. Maybe give us a little more color on OpEx and gross margin trends and how you're going to hit the operating margin target.
Yeah. First, gross margin. The qualitative context is we expect gross margin to come down modestly off of Q4 levels, and a little bit more off full year 2018 levels. And we do expect that that will be largely, although not completely, offset by spending as a percentage of revenue coming down again in 2019. I think just on the gross margin, the trends are going to be a little bit similar to what you've seen in the past, although we do expect a little bit less ASP gross margin improvement from ASP. We expect unit costs to be up a little bit, and that'll be primarily as we ramp 10-nanometer, and then the mix dynamics of more memory and more modem will weigh on gross margin a bit. Year-over-year, we expect gross margins to come down a little bit.
On spending, as you know, we exit this year with spending levels down in the 26% in the Q4, and a little under 29% for the full year. We're way ahead of the three-year plans that we laid out a couple years ago, and we feel pretty good about the progress we've made on the spending. We've done it without cutting R&D. During that timeframe, R&D has grown. We've been investing in the right things. Those things are growing faster. As a result, spending has come down 700 points from 2015 levels. As we go into 2019, spending overall, we expect to come down. Some things we did during the course of the H2 of 2018, including we exited Wind River, we exited wearables, we exited some of our new technology, small, little businesses.
We exited those businesses in the H2 of the year. We did some restructuring in the H2 of the year. As we go into 2018, all that benefit from a relatively low Q4, $4.9 billion run rate, we expect that spending for the full year will be down year-over-year. You net all that together and we've been really focused on growing the operating income dollars of the company. We focus on, but we're not preoccupied with where the gross margin's going to land. Our focus has been on how do we grow the operating income dollars of the company. In a relatively small growth year, we see keeping operating margins at 34% to be a relatively good place without cheating the investments we need to make to continue to progress into 2019 and 2020.
Thank you. Our next question comes from Ross Seymore with Deutsche Bank. Your line is now open.
Hi, thanks for letting me ask a question. Just wanted to follow up, Bob, you gave a lot of great detail there on the margin side, especially on the OpEx side. I wanted to go right back to the gross margin side, though, somewhat simplistically perhaps, you kept the gross margin guide basically the same as you did at the end of last quarter, despite the headwinds to mix, seemingly, with your data-centric commentary, and data center being worse and your revenue being lower. Is there any more color you can give on the puts and takes that leads to just the modest decrease, given those other variables that seem like they have increased as headwinds from when you last talked about gross margin in 2019?
Yeah, you know the gross margin, Ross, isn't really any different. The puts and takes back then, as I indicated, were modest ASP growth. As we are going to fight to protect our market share position, we don't expect a lot of ASP growth. Again, 10-nanometer ramp, not really any different. I highlighted in the prepared remarks, we feel very good about where we are in ramping 10-nanometer during the course of the year to get systems on the shelf for the holiday season. No real change there. Modem and memory growth will be a little bit slower today versus where we were 90 days ago. On the operating margin percent, that's a slight positive. The real only change from 90 days ago is just we're a little more cautious on our revenue outlook. Our spending hasn't really changed.
So we got slight, not as much leverage that we expected back in October, but still good spending leverage during the course of the year. So not really any difference on the gross margin and spending dynamics that we thought 90 days ago, except a little less leverage on the spending line.
Thank you. Our next question comes from Vivek Arya with Bank of America, Merrill Lynch. Your line is now open.
Thanks for taking my question. Within DCG, how should we think about the mix Cloud versus Coms versus Enterprise for Q1 and 2019? Thank you.
Thanks, Vivek. It's Navin here. Look, we've been, over the last 18 months, been working hard to diversify the end customer segment mix inside of DCG. There's three large components, the Enterprise and Government segment, the Cloud segment, and the Coms segment. Cloud and Coms is about two-thirds of the business now, where it was about one-third several years ago. I don't see any major changes to the way things play out in terms of where the growth will come from as we look into 2019 and beyond. While the H1 in the Cloud will be a little bit tougher, we do expect that Cloud continues to grow as they start to move into build out again in the H2. The Coms segment, we continue to gain share in that segment, a large TAM where we have relatively small share.
As we grow our network SoC portfolio, as the market moves to 5G, we expect to continue to gain share. In the Enterprise and Government segment, while we've seen stabilization there over the last four or five quarters, we're not really counting on the Enterprise and Government segment for growth. We do expect that enterprises will continue to make strategic choices about what to deploy on premise and what to deploy in the Cloud. In general, that business is not one that we count on for growth. So in general, you'll see us continue to push on Coms and Cloud to drive growth, particularly in the H2 of the year.
Thank you. Our next question comes from Chris Caso with Raymond James. Your line is now open.
Yes, thanks for letting me ask the question. Just wanted to receive an update on some of the CPU shortages that you've been experiencing, how you're progressing on alleviating those shortages, what effect that may have had on the Q4 results, given that, I guess there was some supply tightness at least coming into the quarter. On that, with demand slowing a bit, is there any fear? Do you have any visibility about customers who may have attempted or succeeded in building some inventory amid those shortages?
First, in the Q4, just in terms of isolating how we prioritize our capacity, server, no shortages. Within the client business, prioritization of big core and to a lesser extent, small core, lower value-oriented product. We do feel like we constrained a fairly healthy PC ecosystem in the Q4. I think when the dust settles on PC TAM, our expectation is it was probably flat and our shipments were down 2%. That was a function of, we delivered every product that we could right up through December 31st. We did have some constraints on the ecosystem and on our customers during the course of the quarter. At the end of the year, Chris, I think inventory levels, relative to the beginning of the year, were a little bit higher, maybe a week and a half, two weeks higher as we enter the year.
Our expectations for the year is the PC TAM is going to be relatively flat, and that for us is a good place to be. We're probably halfway through the PC refresh cycle. We feel things will be relatively flat during the year. Inventory levels in the channels a little bit higher ending the year. For us, inventory levels were relatively low, as you might imagine, on CPU, just because of the constraints we've been dealing with. Our expectation is working with our customers, that we will be through the supply constraints as we exit the Q2 of the year. Again, we'll use the same prioritization of server, big core, small core. We'll be a little bit short on some product mix and on small core until we get probably through the Q2.
That'll constrain us a little bit on just overall growth in the H1 of the year.
Thank you. Our next question comes from Harlan Sur, J.P. Morgan. Your line is now open.
Good afternoon, and thanks for taking my question. Just wanted to get an update on 10-nanometer manufacturability. I know last quarter the team mentioned that 10-nanometer yields were tracking 14-nanometer yields at a similar point prior to production ramp. Is the team still seeing good improvements in 10-nanometer yields? Are you still tracking 14-nanometer yield ramps? Can you just give us an update on early seven-nanometer development and manufacturability?
Harlan, hi. This is Venkata I'll take that question. I can only add to what Bob said in his opening statements, that we continue to make solid progress against our plan that we shared with you during the course of 2018. As I said on the last call, I feel better about our traction today than I did 90 days ago. That continues to bode well for our product launch ambitions, which Bob summarized as having systems on shelf for holiday season 2019, with a barrage of products across all of our businesses to follow shortly thereafter. And I would like to take the opportunity to just remind everybody that at CES, and in the analysts meeting we had at the end of last year, we did show 10 nanometers across the entire portfolio of our product ranges.
We talked about Ice Lake clients, which clearly was top of mind in the early discussions, but we also talked about Lakefield. Bob mentioned that as well. Navin talked about 10 nanometers for Ice Lake server, and we also talked about 10 nanometers moving into our networking and 5G program, which we believe is going to be a big growth sector. The story is not just about 10-nanometer yields, but 10 nanometer now being a key part of our entire product portfolio. As I say, I think that coupled with our focus on the pillars of technology that Bob talked about, in my mind, I think puts our product portfolio looking forward in a pretty good position. Net-net, I think 10 nanometers is looking better now than at the last earnings call.
It's broadly deployed across our portfolio, that in combination with the other technology ingredients that Bob talked about, we believe sets us up for a pretty exciting product roadmap.
Thank you.
Thank you.
Murthy, I think we've got time for two more questions.
Our next question comes from Ambrish Srivastava with BMO. Your line is now open.
Bob, I just wanted to go back to the NSG and the profitability in what was a really booming year for memory. Obviously, prices started to come down back half of the year, NSG was barely profitable. Just from a CFO perspective, what is your tolerance level for having a segment that could go into a deep cyclical downturn, it is heading into a deep cyclical downturn. How do you think about having a commodity within the Intel umbrella and again, the tolerance for lack of profitability? Just your perspective on that, please. Thank you.
Yeah, Ambrish. First, a couple things. I mentioned this in the prepared remarks, Murthy touched on it. When we look at the technologies that we believe are going to be imperatives going forward in this increasingly data-centric world, process, CPU architecture, interconnect, software, memory is a key component, and all the advancements in CPUs will be constrained if you don't have differentiated technology in memory. So we think that the role memory plays going forward is increasingly important. In terms of just the CFO lens of having a commodity in the portfolio, I'm not too excited about it. That's why the investments we're making in memory are for what we believe differentiated technology, both in the manufacturing process capabilities of 3D NAND, but also the differentiated technology for Optane and the role that it plays both on the PC side, but most importantly for us on the data-centric side.
So we're not particularly excited about commodities. When we make these investments, it's really geared towards products and technologies that are increasingly important, and those technologies that are differentiated from kind of the core memory space that help us, in conjunction with the CPU, solve customers' problems.
I'll just maybe add one thing, it's Navin, as an example of that. The Optane persistent memory combined with Cascade Lake, Xeon plus Optane, that is a platform play. Optane persistent memory works uniquely with Xeon. As I think about and talk to customers about the massive amount of data growth we're seeing, the ability for us to uniquely tie those two assets together to solve customer problems is a differentiator for us and allows us to drive growth. To the extent we can exploit more of those kinds of opportunities, things get more exciting from a business unit GM point of view.
Thank you. Our next question comes from Timothy Arcuri with UBS. Your line is now open.
Thank you. Navin, I had a question for you. There seems to be a little bit of a different tone between what we hear from your cloud customers. The compute guys, you, and the memory guys are seeing weakness, but the networking companies still sound fine. Is it just an inventory digestion of computer and of servers, or is there something structural happening there? Thank you.
Yeah. I think as Bob said, I think we sort of talked about a little bit, we had three quarters of really, really strong growth in 2018 in the cloud, that was driven by a product cycle as well as a typical multi-year build-out pattern with Xeon Scalable. If you look back at all the historical trends we've had in the cloud business, we've always said there's some lumpiness to the business, there's periods where people build, there's periods where people consume. The signals we get from our customers is period of build for compute is going to shift now to a period of consumption, that started in the H2 of the Q4, we expect that to continue through the H1 of the year.
Secularly, over the long term, maybe in the long term, the cloud business is going to continue to grow. There's no doubt about that. Both the consumer cloud and the enterprise commercial cloud, we see both of those continue to grow, and the appetite for compute, I think is somewhat insatiable. Bob talked about compute cycle growth. Our five-year forecast for compute cycle growth or MIPS growth is 50% CAGR over the next five years. I see nothing slowing that down over the next number of years. That'd be how I kind of answer that one, Timothy.
All right.
Hey, maybe if I could just close out, Mark. Look, we think 2018 was a great year. Strategically, what it is we're trying to do and the opportunities we see are as strong, if not stronger today heading forward as they've ever been. We think for 2019 for us is going to be another record year. At the same time, we realize that Q1 is just going to be lower, and the practical reality is we think we have a reasonably good read on the level of inventory that's in the ecosystem. I say this particularly for we're getting better and better on the diagnostics around the DCG business. The Q4 to Q1 dynamics for DCG historically have been sequentially down 8% to 10%. The practical reality is, as we see it now, is that could be double in the Q1.
But that has nothing to do with the strength of the business, the product line portfolio we have coming, and our excitement about delivering a real strong 2019 as we go forward. So thank you very much for joining us, and I'm sure we'll talk to you soon.
Operator, please go ahead and wrap up the call.
Ladies and gentlemen, thank you for participating in today's conference. This concludes today's program, and you may all disconnect. Everyone, have a wonderful day.