Good day, ladies and gentlemen, and welcome to the Intel Corporation Q4 2017 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 be given at that time. If anyone should require operator assistance during today's conference, please press the star then zero key on your touchtone telephone. As a reminder, this conference call is being recorded. I would now like to turn the conference over to Mark Henninger, Head of Investor Relations. Sir, you may begin.
Thank you, operator, and welcome everyone to Intel's fourth quarter 2017 earnings conference call. By now, you should have received a copy of our earnings release and the CFO earnings presentation, which replaces the CFO commentary that we've previously used. If you've not received both documents, they're available on our investor website, intc.com. The CFO earnings presentation is also available via the webcast window for those joining us online. I'm joined today by Brian Krzanich, our CEO, and Bob Swan, our Chief Financial Officer. In a moment, we'll hear brief remarks from both of them, followed by the Q&A. Before we begin, let me remind everyone that today's discussions contain 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. CFO commentary and earnings release available on intc.com include the full GAAP and non-GAAP reconciliations. With that, let me hand it over to Brian.
Thanks, Mark. 2017 was a record year for Intel, and fourth quarter results were outstanding, well ahead of the forecasts we outlined in October, based on the strength of both our PC-centric and data-centric businesses. I'll review our results with you in just a moment. Before I do that, I'd like to share a few words about security. We've been working around the clock with our customers and partners to address the security vulnerabilities known as Spectre and Meltdown. While we've made progress, I'm acutely aware that we have more to do. We've committed to being transparent, keeping our customers and owners appraised of our progress, and through our actions, building trust. Security is a top priority for Intel. It's foundational to our products, and it's critical to the success of our data-centric strategy.
Our near-term focus is on delivering high-quality mitigations to protect our customers' infrastructure from these exploits. We're working to incorporate silicon-based changes to future products that will directly address the Spectre and Meltdown threats in hardware, and those products will begin appearing later this year. However, these circumstances are highly dynamic, and we've updated our risk factors to reflect both the evolving nature of these specific threats and mitigations, as well as the security challenge more broadly. Security has always been a priority for us, and these events reinforce our continuous mission to develop the world's most secure products. This will be an ongoing journey, but we are committed to the task, and I'm confident we're up to the challenge. To keep you informed, we've created a dedicated website, and we're approaching this work with customer-first urgency.
I've assigned some of the very best minds at Intel to work through this, and we're making progress. With that, let's turn to our 2017 and fourth quarter results. We just wrapped up the best year in Intel's history with the best quarter in Intel's history. Revenue was up 4% year-over-year in the fourth quarter, 8% if you exclude McAfee, setting an all-time record. Our data center, IoT, and FPGA businesses each set revenue records. We met or exceeded all of the financial commitments we made to you at the beginning of the year, and our focus on efficiency and profitable growth produced significant leverage, driving non-GAAP operating income up 21%. Our data-centric businesses deliver the technology foundations for the new data economy, making the analysis, storage, and transfer of data possible, giving our customers the ability to turn data into amazing experiences and actionable insights.
They are central to our strategy. Data-centric revenue, excluding McAfee, was up an impressive 21% over the fourth quarter of last year. I'd like to share a few highlights with you, starting with DCG. DCG's revenue was up 20% over the fourth quarter. The cloud segment was up 35%, comm service providers were up 16%, enterprise was up 11%, and our adjacency revenue was up 35%. We saw broad-based demand strength with customer preference for high-performance products driving richer ASP mix. Cloud segment results were driven by significant volume growth and continued customer preference for higher performance products. In the comm service provider segment, we continued to take share and grow revenue as customers chose IA-based solutions to virtualize and transform their networks. Enterprise segment strength was driven mostly by ASP as customers transitioned to Xeon Scalable products in a seasonally strong fourth quarter IT buying window.
While we continue to see enterprise customers offload workloads to public clouds, we are also seeing those customers prioritize performance solutions for hybrid and on-premise build-outs. Customers across all of our segments are accelerating deployment of Xeon Scalable processors, which is ramping roughly in line with our historical Xeon transition. We continue to demonstrate leadership and progress in artificial intelligence with the data center to the edge, and from hundreds of watts to milliwatts. Our software optimization for the Caffe framework has improved already strong Xeon Scalable ResNet-50 inference performance by two times just since the launch in July. The first-generation Nervana neural network processor ran a neural network less than two weeks after we received silicon, and we've shipped our first customer units.
At the edge, Google announced its AIY Vision Kit, featuring our Movidius vision processing unit, and Amazon announced DeepLens, the world's first deep learning-enabled video camera for developers, which uses an Intel CPU, graphics, and compute libraries for deep neural networks. We're also seeing design wins that combine technologies from multiple data-centric business units, reinforcing the idea that we have developed a unique and differentiated collection of capabilities that can address customer challenges together more effectively than any one business could alone. A great example is Dahua's recent announcement of their DeepSense product line, which combines core CPU, Intel FPGA-based network video recorders, along with Movidius VPU-based cameras to enable people and automobile detection in smart city applications using artificial intelligence. In the Programmable Solutions Group, we saw strong double-digit growth in the data center, auto, embedded, and advanced products categories, as well as last-time buys of legacy products.
That strength was partially offset by softness in the comms infrastructure. In Q4, we launched the Intel FPGA SDK for OpenCL to dramatically increase productivity for our customers. We also delivered first-to-market leadership innovations in the Stratix 10 product line, including the first SoC FPGA with an Arm processor at more than one million logic elements and the industry's first FPGA with integrated HBM2 memory. Our Internet of Things Group grew 21%, with continuing momentum in retail, video, and in-vehicle infotainment verticals. Wind River saw strong multi-year contract renewals and delivered its most profitable quarter ever for Intel. The memory business grew 9% and achieved profitability in Q4. A strong client volume was partially offset by a one-quarter qualification delay of a data center's SSD volume. Last quarter, I shared with you that our leadership technology is resulting in strong customer interest in long-term supply arrangements.
That interest has continued to grow. We have since signed additional agreements and now expect pre-paid payments totaling roughly $2 billion over the course of 2018. Mobileye had another strong quarter, and business momentum is growing. Our 30 design wins over the course of 2017 and 15 new program launches in 2018 are both increases of two and a half times over the prior period. We now have level 2 plus and level 3 design wins with 11 automakers, who collectively represent more than 50% of global vehicle production. These advanced programs launch over the next two years, and they represent a major leap in functionality versus current semi-autonomous systems and are a significant step towards a scalable level 4 and level 5 fully autonomous system.
We reached an important milestone in the fourth quarter, the announcement of our level 3 through five autonomous driving platform based on EyeQ5 and Atom, which we'll sample over the next few months. We believe this will be the most advanced, scalable, and efficient platform of its kind. EyeQ5 will deliver 24 teraops of deep learning performance in a 10-watt power envelope or about two and a half times the efficiency of the competition. Just a couple of weeks ago at CES, we announced that by the end of 2018, we expect two million EyeQ-equipped cars will be collecting crowdsourced data for REM, our Road Experience Management mapping solution. The resulting map will first be utilized in level 3 beginning in 2019.
The ability to crowdsource data to build and rapidly update the precision maps required for higher levels of automation is a major differentiator in our plan to build out the safest and most affordable autonomous vehicle system. It's also a great example of our data-centric strategy at work. Finally, I'd like to touch on our PC-centric businesses, the Client Computing Group. Over the course of the year, the PC market improved. Our 14-nanometer manufacturing cost came down, and the competitive environment intensified. Against that backdrop, PCG's focus on innovation and performance, especially in growth segments like gaming, two-in-ones, thin and light notebooks, and enterprise, led to record core mix and record i7 volume in the fourth quarter. We also shipped our first low-volume 10-nanometer SKU, and our modem business grew 26% over the fourth quarter of last year.
Intel is undergoing one of the most significant transformations in corporate history. From a PC-centric company to a data-centric company. We've made thoughtful, disciplined investments along the way that have expanded our TAM to $260 billion. Those same investments have produced a collection of data-centric businesses that are unmatched, growing at double-digit rates, and approaching 50% of the company's revenue. Our opportunity is larger than it's ever been, and we're hungry to compete and win. In 2018, our highest priorities will be executing to our strategy and meeting the commitments we make to our owners and our customers. This includes our commitment to restoring customer confidence in the security of their data. This year, Intel will celebrate a half-century of innovation that has profoundly changed the world.
Over the last 50 years, we invented the architecture and manufacturing technologies that have made personal computing, the internet, and the cloud not only possible, but pervasive. The journey hasn't been without challenges. Nothing worth doing ever is. Our culture has been forged through taking challenges head-on and developing solutions our customers can count on. That includes working directly to address the Spectre and Meltdown security threats. We leave 2017 on a financial high note, but I'm even more excited about what's to come, about our strategy producing great products for our customers and great returns for our owners. I see Intel innovation changing the world for another 50 years, and that journey starts with 2018.
Over the coming year, we'll bring amazing innovation and performance to the PC market, advance the state of art in artificial intelligence, lead the way towards mass 5G deployment, launch the industry's first new memory architecture in two decades, and take another step toward a safer world in which autonomous driving is a reality. Looking back on 2017, I could not be more proud of our team and all they have accomplished. As I look to our 50th year, I'm more optimistic and confident than I've ever been about Intel's future. With that, let me hand it over to Bob.
Thanks, Brian. The fourth quarter was an outstanding close to a record 2017, and we are building real momentum heading into 2018. Revenue for the quarter was $17.1 billion, up 8% year-over-year. Operating income was $5.9 billion, up 21% year-over-year, and EPS of $1.08 was up 37% year-over-year. From a capital allocation perspective, we redeemed $1.6 billion of 2035 convertible debt, reducing our share count by 59 million shares, and repurchased $500 million of higher coupon debt and exchanged $1.8 billion into 30-year debt at a 1% lower coupon rate. On tax reform, our Q4 GAAP earnings reflect a one-time tax impact of $5.4 billion. Our guidance reflects approximately seven-point improvement in our effective tax rate going forward.
To summarize, we had a fantastic quarter and year and are on track to exceed the three-year targets we laid out at our Analyst Day one full year ahead of schedule. Our Q4 results demonstrated continued momentum in our transformation from a PC-centric to a data-centric company. Intel's data-centric businesses, those outside of the PC segment, are at an all-time high mix of 47% of our revenue, up from approximately 40% in 2012. We've made significant investments to expand our TAM into new data-rich markets like memory, programmable solutions, and autonomous driving. These investments are just starting to pay off and will fuel Intel's growth going forward. Our PC-centric business was down 2% in a declining PC market, it continues to be a great source of profitability. DCG delivered its most profitable year since 2011 by focusing on premium and growth segments with industry-leading products.
This business generated the cash to fund Intel's investments in new data-centric growth. Moving to Q4 earnings, we generated significant EPS expansion in the quarter, up 37% year-on-year. Our EPS improvement was driven by strong top-line growth, a five-point improvement in operating margins, and significant gains from our ICAP portfolio. Our gross margins expanded two points in the quarter, our spending as a percent of revenue declined by three points as we delivered $700 million more revenue on $300 million less spending. In terms of operating efficiency, we're well ahead of schedule of meeting our commitment by reducing spending to 30% by 2020. We now expect to achieve this goal no later than 2019. Total spending was down 6% year-over-year in the fourth quarter while we continued investing in our key priorities, including driving Moore's Law forward, winning in artificial intelligence, and autonomous driving.
R&D spending as a percent of revenue was down approximately one point, our SG&A costs were down over two points as we rationalize our marketing and sales programs and generate significant leverage in our SG&A functions. Let me touch briefly on our segment performance. Client Computing Group had another strong quarter. Revenue of $9 billion was down two points, operating margins were down two points. Operating margins were lower on 10-nanometer transition costs. We saw strength in the commercial gaming business, we believe the worldwide PC supply chain is operating at healthy levels. The Data Center Group had record revenue of $5.6 billion, up 20% year-over-year, operating income of $3 billion, grew 59%. Q4 operating margin was 54%. As Brian mentioned earlier, we had strong growth and execution across all segments.
Overall unit volume was up 10%, ASPs were up 8%, our adjacencies grew by 35%. Our ASP strength demonstrates the value customers see in our high-performance products. Xeon Scalable, launched in July, is ramping well with customers broadly deploying this leadership product family. Revenue scale from leadership products, ASP strength, and the exclusion of 2016 one-time charges drove strong operating income growth for the business. For the full year, revenue was up 11%, operating margin came in at 44%, both ahead of the expectations we provided at the beginning of the year. The IoT, NSG, and PSG business segments are becoming a larger component of our overall business, collectively growing 19% year-over-year. Our Internet of Things business achieved record revenue of $879 million, growing 21% year-over-year, driven by strength in industrial and video and continued momentum in our retail business. Operating profit was $260 million, up 43% year-over-year.
The Mobileye business also had a record quarter, we are on track to our deal thesis. As we called out last quarter, results for the Mobileye acquisition are included in our All Other segment, reflects the Q4 integration of Intel's autonomous driving group spending into Mobileye. Our memory business had revenue of $889 million, up 9% year-over-year, with strong demand for data center SSD solutions and demand signals outpacing supply. This segment was profitable for the quarter, we expect the segment to be profitable for the full year of 2018. Programmable Solutions Group had record revenue of $568 million, with 35% growth driven by strength in data center, automotive, embedded. Operating profit was $156 million, up 95% year-over-year.
The Stratix 10 design win pipeline, which represents PSG's largest ever, doubled over the last year due to engagements in 5G, cloud computing, the infrastructure transition to network function for virtualization. We laid out our capital allocation priorities early in the year. Invest organically, expand acquisitively, return capital to our shareholders, do it wisely. In the year, we delivered on our promise. First, we generated strong free cash flow of $10.3 billion in the year, returned $8.7 billion to shareholders through dividends of $5.1 billion and share repurchases of $3.6 billion. Second, we funded a majority of the Mobileye acquisition from the sale of non-core assets during the year, including McAfee and the sale of ASML shares. Third, we redeemed $1.6 billion in convertible debt, reducing 59 million shares. We also tendered higher coupon debt for lower coupon debt.
Let me expand on the ICAP and treasury-related transactions we executed in the quarter. First, we sold 11.4 million shares of our ASML holding, which generated $2 billion in cash proceeds and a gain of $1.5 billion. Second, we redeemed our 2035 convertible debenture. This redemption created a $2.8 billion cash outflow and a non-cash loss of $385 million, which was tax-deductible. Since this debenture was convertible, therefore dilutive, the redemption effectively acted as a buyback that will reduce diluted share count by 59 million shares. Third, we successfully tendered $2.3 billion of high coupon debt in the quarter. We exchanged $1.9 billion of old debt into $2 billion of 30-year new debt, reducing our coupon rate by 1%. We also redeemed $425 million of old debt for cash. This transaction both lowered our leverage and our interest expense.
Adding it all up, 2017 was another record year for Intel. Revenue of $62.8 billion was up 9% year-over-year, driven by 16% growth in our data-centric businesses and 3% growth in PC-centric business. Operating income of $19.6 billion was up 18% on strong execution across the businesses and disciplined spending. Earnings per share of $3.46 was up 28% on excellent operational performance and the benefit of $0.35 from ICAP net gains. With the change to the accounting rules for recognizing price changes on equity investments, we do not expect to see these ICAP gains repeat. Before I turn to guidance, let me provide a little more context on the impact of tax reform on our business. Intel's fourth quarter results reflect a higher GAAP income tax expense of $5.4 billion as a result of U.S. corporate tax reform enacted in December.
This includes a one-time required tax adjustment on previously untaxed foreign earnings, payable over eight years, which was partially offset by the remeasurement of deferred income taxes to the new U.S. statutory tax rate. Looking ahead, we expect the Tax Cuts and Jobs Act will help level the playing field for U.S. manufacturers like Intel that compete in today's global economy. We expect a 2018 tax rate of approximately 14%, driven by a lower U.S. statutory tax rate of 21%, lower tax on foreign income, benefits from U.S. exporters, and the continuation of the R&D credit. The change in our tax rate drives approximately $0.28 in 2018 EPS. Intel has a rich history of investing in U.S.-led research and development and U.S. manufacturing. Just last year, we committed to the fit-up of our Fab 42 facility, creating thousands of jobs at completion.
These tax reforms provide further incentive to continue investments like this. Before we turn to our 2018 outlook, I also want to highlight two accounting rule changes. First, new accounting rules for revenue recognition, second, accounting for equity gains and losses. We do not expect a material impact to revenue from the revenue recognition accounting change. The changes in accounting for equity gains and losses will require the recognition of unrealized price changes each quarter. Equity holdings like our ASML position will see mark-to-market adjustments that will flow through earnings in 2018, which may create greater volatility on a GAAP basis. This change resulted in an impact of $2.7 billion of net unrealized gains at year-end that we booked to equity on January 1st, 2018. Now, moving to the full year. We are forecasting the midpoint of the revenue range at $65 billion, up 4% year-over-year.
We expect operating margin of approximately 30%, with gross margins down 2-2.5 points and spending as a % of revenue to be down 1-1.5 points. The decline in gross margin is driven by growth in our adjacent businesses as we play in an expanded TAM, and transition costs associated with 10-nanometer, both partially offset by higher gross margins from our 14-nanometer products. We expect EPS of $3.55, up 14%, excluding the ICAP net gains, driven by a strong top-line growth and a lower tax rate of approximately 14%, which will increase EPS by approximately $0.28. We expect net capital deployed of $12 billion. This reflects gross CapEx of $14 billion, offset by approximately $2 billion of customer prepayments for memory supply agreements. The increase in CapEx reflects our and our customers' confidence in our memory technology leadership.
We expect free cash flow of $13 billion, an increase of approximately 30%, directly contributing to our decision to raise our dividend by a full 10%. As we look to the first quarter of 2018, we are forecasting a midpoint of revenue range at $15 billion, up 5% year-over-year. We expect operating margin of approximately 27%, flat year-over-year, with a 3-point decline in gross margins offset by a 3-point decline in spending. We expect EPS of $0.70, up 11%, excluding ICAP net gains, from strong top-line growth and a lower effective tax rate. We believe 2018 will be another record year for Intel, and we feel great about where we are entering year two of our three-year transformation. We've met and exceeded our commitments.
Our PC-centric team continues to operate very well in a down market. Our data-centric businesses are up double digits collectively as we continue to transform the company to power the cloud and smart connected devices. With that, let me turn it back to Mark.
All right. Thank you, Brian and Bob. Moving on now to the Q&A. As is our normal practice, we would ask each participant to ask one question and just one follow-up if you have one. Operator, please go ahead and introduce our first questioner.
Our first question comes from the line of John Pitzer from Credit Suisse. Your line is now open.
Yeah. Good afternoon, guys. Congratulations on the strong results. Thanks for letting me ask the question. My first question just revolves around the revenue guidance for the March quarter. At the midpoint, down about 12% sequentially is a lot worse than, I guess, normal seasonal. I guess, I'm trying to figure out kind of the parameters that you guys are using to come up with that number. Was there something in the DCG strength in the calendar first quarter that you don't think is repeatable? Given your comments around the PC supply chain being healthy, that to me feels a little bit more seasonal than not in the March quarter. I'm just kind of curious as to why the Q1 revenue would be so much below what has been kind of the five-year medium seasonality for Q1.
John, thanks. It's Bob. Really two things going on. One, the enterprise growth in DCG, as Brian highlighted, was up 11% in the fourth quarter, extremely strong seasonal growth for enterprise, a little bit more than we expected, frankly. That's number one. Number two, PSG, 35% growth was helped by end-of-life sales during the course of the quarter. If you adjust for maybe a more normalized enterprise growth and PSG growth, you get to more of a seasonal Q4 to Q1 dynamic. In line with seasonal, if you make those two adjustments.
That's very helpful. Brian, on my second question, it revolves around CapEx. I think it was a year ago on this conference call that you kind of mentioned that you thought that calendar year 2017 and calendar year 2018 would be sort of above trend line CapEx spending for you guys, and then you thought it would come down again in calendar year 2019 to a more normal trend, albeit you didn't quantify the trend. Just given the big uptick in 2018 to $14 billion, I realize two of that's already being covered by prepayments, but how do we think about kind of the trend line CapEx from here, especially in light of the changing relationship between yourself and Micron on IMFT?
Sure, John. My perspective is, if you take a look at it, right, logic on our CapEx just scaled with our increase in revenue and kind of our overall growth rate. That's felt in line. With memory, we're going to take a look at it really on a year-by-year basis. I'd tell you that Bob's really doing a good job of helping the business unit look over the capital. When we have demand and people are willing to pay up front for that demand and reserve the capacity like we've done this year, we're going to go ahead and put that capacity in place, plus what we think the overall market, we can do as distributing across the overall market. That's really independent of the changing relationship with Micron.
That's really more about how we're doing development work out in time, really, and that's actually still two generations away. That didn't really affect 2018. This was really about what we saw for the overall NAND memory market plus the additional capital and capacity that people wanted to reserve through that process. We'll look at that each year, John, and say, okay, as we look out into 2019, at the end of 2018, we'll do that same analysis.
Awesome. Thanks.
Sorry, I was just follow on to Brian Krzanich's comments. The way we're looking at memory, and we talked a bit about this on the last call, is increased confidence in our customers in the technologies that we're developing, where those relationships, they will help fund the scaling of the capacity to grow the business. We're really trying to match net capital employed to be in conjunction with known customer demand. What that meant for 2017 is the net capital was roughly $1.5 billion. We've held that relatively flat in 2018. While gross capital is higher, we have more conviction in the customer base to fund $2 billion of the gross capital. Memory, we're really trying to focus on customer adoption of our technology to effectively and efficiently scale the business.
Just the only other point I would make on logic, remember, a year ago, in terms of our outlook for growth, we're well ahead of our outlook for growth. We were probably at the time implied in our outlook was probably a $62.5 billion kind of number in 2018. Obviously, with our guide now at $65 billion, we're $2.5 billion higher, it's that incremental growth that's placing more demands on logic capacity, both for 14, 10 nanometer N as we think forward about seven nanometer.
Perfect. Thanks, guys.
Thank you. Our next question comes from Joe Moore from Morgan Stanley. Your line is now open.
Great. Thank you. I wonder if you could talk about OpEx. Obviously, you've been pretty disciplined there, bringing that down, but you've also got some other initiatives. You've announced a discrete graphics effort. Maybe if you talk about, it seems like that would cost a lot if I just look at what your competitors are spending. At some point, you're going to spend money on NAND that used to be shared with Micron. Can you talk about the puts and takes there, and is there something that we should think about that's coming down to sort of offset those potential increases?
Sure. I'll start and then I'll let Bob kind of give you the under-the-cover detail of the dollars. Joe, really, we've already factored all of those things in. Things like the discrete graphics is a ramping spend. The memory is R&D spending out in time, so for 2018, has no really effect. We've driven an overall efficiency in all of our R&D spending to offset that. Increasing in GPU spending. There's some other increases as well around things like autonomous driving and some of the other artificial intelligence in some of the emerging areas. You're right. Over time, we'll increase spending in NAND, in R&D, those are being offset by efficiencies that we're driving into the rest of our product R&D.
We really feel like we're getting to a good point where we can keep the pace of innovation going on our core products while we fund these new initiatives as well, and not take a beat from our continued efficiency efforts across the spending as a % of revenue.
Joe, I'll just add some numbers to Brian's words. We've kind of come down from 36% to 35% to 34% to a second half of 2017 at roughly 31% of revenue. We've been coming down as we've been doing two things. One, the investments that we've been making are paying off in terms of higher growth. Number two, we're making real trade-offs in where we're investing our money. Implied in our guidance as we go into 2018, we're expecting our spending level to be roughly flat with an annualized 2017 spending level. Underneath that, continuing to drive efficiencies and sales and marketing as we become more of a B2B or data-centric company and getting real leverage on our G&A functions across the board. We've made real progress during the course of the last couple of years, including the second half.
We expect to make continued progress while making the critical investments in things like discrete graphics, autonomous driving, artificial intelligence, and continuing to invest in Moore's Law.
Okay, great. As you think about that NAND investment, I guess people have asked me what the separation from Micron on the long-term path means. Does that mean there's sort of more of a focus on proprietary products like 3D XPoint, or do you remain firmly committed to a more competitive NAND market?
Sure. Just want to make sure we clarify a little bit, Joe. The separation of development work is, again, out in time. Think of it in the 2020 timeframe is when the real independence comes, and it's NAND specific. We continue to work together on 3D XPoint. This is really just, it's not a separation of the companies or something about the relationship. The relationship with Micron continues to be a good one. I foresee it will continue to be a good one in the future as well. This is about direction of where we're going to take our products, and we've talked about ours are very data center centric, and really tied to performance and aligning to the customer market that we're really looking at. The spending in R&D we're talking about right now is purely NAND specific.
3D XPoint continues to be a joint effort.
Understood. Thank you very much.
Thanks, Joe.
Thank you. Our next question comes from Ambrish Srivastava with BMO Capital Markets. Your line is now open.
Hi. Thank you very much. Brian and Bob, I wanted to go back to DCG, specifically on the op margin front. We have not seen a five handle, if my model is correct. It's been eight-plus quarters. Can you please speak to the sustainability of the op margin and kind of what were the drivers that got you to that level? I had a follow-up as well, please. Thanks.
Yeah. We came into the year, and we indicated that we expected margins for the full year to be in the kind of 40%-45% range. Obviously, we started out low, but we set our expectations were that would grow throughout the course of the year. A couple of things in the fourth quarter. Obviously, growth. Real good leverage on our existing investment was a big contributor. Secondly, ASPs were up 8%. Of the 20% growth, ASPs accounted for eight points of that. As Brian highlighted, whether it's cloud, comms, or enterprise, customers in the quarter were really paying for performance. That performance for us was higher ASPs. Third, continued progress on unit costs.
Last, you may remember, last year's fourth quarter, we had some warranty and IP-related charges in the data center business that I think cost us roughly four points a year ago. Last year's were a little bit deflated. Good volume leverage, strong ASPs as customers paid for performance, and unit costs improving.
We should expect this level going forward, Bob?
I think I wouldn't get too far away from the 40%-45%, to be honest with you. I think as we go into 2018, we think good cloud momentum, that's been consistent performance over the course of the year, last couple of years. Coms, our performance, we believe, was real strong in a somewhat sluggish market, so real share gains. The high kind of seasonal enterprise growth and strong ASPs in the quarter, we think are more seasonal in nature, and we're not anticipating enterprise growth to stay at these levels as we go into 2018.
Okay, that's helpful. Then for my follow-up on the gross margin for next year, you provided the sort of qualitatively, the adjacency as well as the cost impact from the 10nm. A, I was surprised you didn't mention competition because now AMD will have a full year of a product in an area they never were for, I shouldn't say never, they were not in an area for a long time. Question is, are you seeing any competition? Are you factoring that in? B, would it be possible to quantify the two impacts that you mentioned? Thank you.
I'd highlight three things that are driving the deterioration. I'd start first with, we've always characterized our long-term gross margins to be in the 55%-65% range. For the last several years, and included in our guidance this year, we'll be at the upper end of that 60%-65% range. For the last couple of years, we've had maturity of 14-nanometer, and that maturity, both in terms of getting more and more performance, and lower and lower unit costs, has been contributors to gross margins. As we go into 2018, we see 14-nanometer modest contributions from profitability because we've matured on the cost curve. It's an increasingly competitive environment, and we don't anticipate dramatic ASP improvements over the entire company. Some cases, yes, but for the most part, no.
We'll see some continued improvement in 14-nanometer products during the course of the year. Secondly, we are really accelerating the growth of our adjacent businesses, the investments that we've been making in both the data-centric businesses and our growth in modem, those are contributing to our year-on-year EPS growth. However, both of those product lines or businesses have lower margins. Our success in modem and memory is having a mixed impact on our gross margins. The third area is, we're going from 10-nanometer startup, where costs are coming down, to more the ramp to the 10-nanometer. That ramp is going to weigh on the margins as we begin to develop production going into the second half of the year, where we're way up the curve on where our yields and costs are.
Those are kind of the three things, as we see it, and if I were to characterize, roughly plus one, roughly minus two, roughly minus one, rounding those.
Ambrish, on your question about specifically around competition, I think every year we look at it as a competitive environment, and we're out to compete for our customers. We factor that in each year appropriately, I think, against the competition. We've looked at the competitive environment and we believe beyond Scalable, the great performance that has our overall product roadmap, we think we have a highly competitive roadmap and have adjusted for that in our forecast.
Okay. Thank you for all the details. Good luck.
Thank you. Our next question comes from Stacy Rasgon from Bernstein. Your line is now open.
Thanks for taking my questions. Can you walk us through your free cash flow waterfall? I can't get there. Flattish operating income on 4% revenue growth. You lose a couple billion dollars in ICAP gains. CapEx, I guess maybe it's flattish with some MP payments. Taxes maybe gets me a third to half of the way there. Where does the rest of the free cash flow come from? Are you just draining a ton of working capital or what? Can you walk us through?
Yeah. $13 billion in free cash flow, up roughly 30% year-on-year. Higher cash earnings. Our guide has EPS up roughly 14%. In that guide, there's depreciation has grown in 2018 relative to 2017. That has just higher cash earnings. Secondly, we do expect lower working capital as we go through 2018. Third, we have higher strategic customer supply agreements. Those three things are all contributing to the positive. Obviously, Stacy, the higher gross capital is a bit of an offset. Stronger cash earnings, better working capital, more strategic supply, partially offset by higher CapEx.
Okay. For my follow-up, I want to ask about the growth next year of your kind of adjacent businesses versus your core businesses. Of that $65 billion, of that 4%, can you give us a feeling how much of that is coming from the adjacent, I guess, mostly memory and modems, versus the core? It must be decent given the gross margin pressures. I guess, as a corollary to that, you mentioned a one quarter delay in data center memory. Was that a statement on Crosspoint?
No, that was a NAND. I'll start with just the answer to that question for you, Stacy. That was a 3D NAND SSD. We had some delays in the fourth quarter, and those are being addressed now. That was that comment.
Okay. Oh, sorry, go ahead.
No, I think on your first question, in the $65 billion, we characterize it as roughly low single-digit decline in our PC-centric businesses. Implied in that is PC maybe declining a little bit more and modem, i.e., adjacency within the CCG segment, partially offsetting that. In the overall guide, we said that the data-centric businesses would be growing in the mid-teens, and obviously we believe that'll be the strongest growth segment within the makeup of our data-centric businesses will be in memory. It will be a function of customer quals that Brian just highlighted, and we expect NSG growth to accelerate throughout the course of 2018.
What is it?
Low single-digit decline on PC-centric, mid-teen growth on data-centric businesses.
What does that memory strength imply for the data center growth, the DCG growth, particularly given the strong performance in Q4?
For the most part, the bins for NSG that will go through data center will be late in 2018 and won't really have an impact on the overall growth rate of that business. It's primarily about growth of 3D NAND during the course of the year. Thank you.
Got it. Thank you.
Thank you. Our next question comes from C.J. Muse from Evercore. Your line is now open.
Thank you for taking my question. I guess, a couple housekeeping questions, if I could kind of put them together. Curious if you could share with us how you're thinking about CapEx spend between logic and memory. Then on the 10 nanometer startup, can you share with us when you're expecting to begin depreciating those costs?
First, on the CapEx, we indicated gross capital in the year of $14 billion, that we'd get customer strategic supply agreements of roughly $2 billion. Our net capital is $12 billion in the year. Again, I kind of break that into two pieces. Memory net capital employed, no change, and logic CapEx up roughly $1 billion year-on-year. As we mentioned earlier, that $1 billion is a function of primarily growth, continuing to grow 14 nanometer. Second, scaling up 10 nanometer, and third, investing in next node 7 nanometer during the course of the year. Those three things are really driving the $1 billion increase in CapEx for logic. The second part of your question.
For 10 nanometer, as we bring that equipment online, there's some equipment now that's being depreciated, but as we bring more online, as we ramp in the second half of the year, when we turn that equipment on is when we start depreciating it. We'd expect our depreciation bill in the second half of the year to be growing.
Thank you.
Really to come full circle, that's one of the contributors to the gross margin a deterioration during the course of the year as we start to ramp 10-nanometer.
Very helpful. Thanks, Bob.
Thanks.
Operator, I think we have time for a couple more questions.
Thank you. Our next question comes from Vivek Arya from Bank of America Merrill Lynch. Your line is now open.
Thanks for taking my question. For the first one, Brian, I'm curious, are you baking in any effect on sales or costs or pricing from any resolution on the processor security issues? There is one line of thinking that says customers might decelerate their purchase, then you have others in the industry saying that customers might accelerate their purchase later on. I'm just curious how you're looking at the financial implications, positive or negative, from this issue, near term and longer term.
Sure. Let me try and kind of break it into two kind of answers for that, Vivek. From a cost standpoint, we've baked in and we've talked about that we don't expect any material impact of this security exploit on our spending or product costs or any of that. That's how we baked that in. From a forecast standpoint, we actually made our forecast, and we've checked it as we go through the first few weeks here of the year against our prior forecasts to make sure that the forecast incorporated any changes or any signs we're seeing up or down. I tell you, at the highest level, we're not seeing much of a change in those forecasts as a result of this. I'd tell you it's pretty balanced right now.
Spending, not material, didn't make any adds there, our forecast, we had a forecast. We've checked it as we go through the first few weeks of the year, it hasn't really changed or altered as we looked at it.
The only other thing I'd add, per Brian's comment earlier, we kind of go into the year realizing that it's an increasingly competitive environment, and our focus is on right now continuing to bring the best, highest performance products to market, but also lots of time and energy spent on focusing on fixing this issue primarily through software patches as opposed to short-term hardware things.
All right. As my follow-up, for the full year, how should we think about growth in just the DCG business? I understand you gave some color around the entire data-centric group, which includes memory and other segments, but just sort of apples to apples, how should we think about growth in just DCG, which had a very strong double-digit growth year in 2017? Thank you.
Yeah. Throughout the course of 2017, we kind of guided to high single digits, we kind of executed to that throughout. Q4 wasn't dramatically different than the first three quarters of the year, really with the exception of the high seasonal spend for enterprise. That really took us from a high single digit to low double digit or 11% growth for the year. We go into 2018, we think that in the enterprise space, things will go back down to the negative single-digit range, therefore don't anticipate a dramatic difference in how we laid out DCG a year ago. That's how we're thinking about it. If I maybe elevate that back up a little bit and just think about when we're putting together our plans for the year, there's two areas where our tendency is to be a little cautious on the outlook.
One is PC TAM. We tend to be a little more cautious on PC TAM, we tend to be a little more cautious as we think about enterprise growth in the DCG business. The reason we do that is we think it's important to be cautious, get our cost in line, and if our assumptions on market rates of growth turn out to be conservative, we believe we'd benefit from higher volume and real strong flow-through to net income. That's how we planned the year coming into 2017, we did kind of the same thing for 2018. I don't anticipate dramatically different DCG growth than how we laid out 2017 as we enter 2018, because we really haven't assumed a change in trajectory of enterprise CIO spending in the course of the year.
Okay. Thanks. Very helpful.
Operator-
Thank you.
If you could go ahead and give us our last question.
Yes. Our last question comes from Blayne Curtis from Barclays. Your line is now open.
Hey, guys, thanks for squeezing me in. I just wanted to go back to the DCG ASPs. You had talked about the ramp of Scalable being kind of to plan, but then I think you got a nice tailwind, and I think you implied enterprise had a higher percent of Scalable. Just wanted to get a talk on broad strokes, just where the Scalable ramp is here and kind of your expectations kind of getting through the year of where that could go.
Sure. I'll start, and Bob can add some number detail and all. The Xeon Scalable ramp is right in line with prior ramps of similar products on our PCG roadmap. What you saw when we talked about Q4 was not necessarily more Xeon Scalable, but people buying up the stack on Xeon Scalable, which drove the ASP. They're buying the higher performance, higher priced parts. That's not uncommon when at the early stages of a ramp, people come in and they typically want to buy the highest performing parts at the beginning, and then they fill out their distribution as other buyers come in and other parts of the market kind of open up. Ramp's on schedule and aligned with prior ramps, and the ASP was more about SKU, like buying up on the higher performance parts than necessarily a volume statement.
Got you. I also want to go back on just Sorry. Thanks. I just want to go back on gross margin as well. As you look for March, you mentioned the drop is from adjacent business as well as 10-nanometer. Obviously, memory, you're signaling a big ramp, but it doesn't move that fully. I'm just wondering if, Bob, you just walk through the 10-nanometer startup costs as they kind of move through the year and if you can outline between those two and Q1, what's the bigger factor?
I think, mixed dynamics are going to be a bigger factor overall for the year, both in Q1 and for the rest of the year, as our memory and modem business continue to accelerate strong growth. That's going to be the biggest impact. I think, 10-nanometer will have an impact just right out of the gate, and will kind of continue throughout the year as we scale volume, but also need to improve yields during the course of the year. Hopefully, that's helpful.
Okay. Thanks, guys.
All right. Thank you all for joining us today. Operator, please go ahead and wrap up the call.
Ladies and gentlemen, thank you for your participation in today's conference. This concludes today's program. You may now disconnect. Everyone, have a great day.