Microprocessor giant Intel (NASDAQ: INTC) has been immensely successful in the market for chips that power major data centers. It has managed to protect its nearly total dominance of the enterprise and cloud server processor markets while growing its share in areas such as networking, storage, and high-performance computing accelerators.
Intel data-center chief Diane Bryant at the company's annual developer forum. Image source: Intel.
Further, with its purchase of Altera last year, Intel further broadened its portfolio of data-center technologies to include FPGAs. Such chips are increasingly being used to accelerate certain data-center workloads; Microsoft (NASDAQ: MSFT) , for example, uses them to accelerate Bing searches .
Intel has built a phenomenal portfolio of technologies to support its excellent processors in the data center. However, there is a risk to this business looming on the horizon that the company needs to be ready to respond to with a relatively early launch of server processors built on its 10-nanometer technology.
The competition is moving to foundry 7-nanometer tech
Intel's main rival in logic chip manufacturing is Taiwan Semiconductor Manufacturing Company (NYSE: TSM) . TSMC's executives have been very vocal about how makers of "high-performance computing" chips are adopting its upcoming 7-nanometer technology.
The use cases that TSMC cites here are "GPU, gaming, PC and tablet, virtual reality, server, FPGA, automotive, and networking applications." TSMC also says these customers "all have aggressive product tape-out plans, starting from [the] early second quarter" of 2017.
Based on this, it would be reasonable to expect that at least one of Intel's server processor competitors will bring a product to market manufactured in TSMC's 7-nanometer technology by the end of 2018.
Intel's 10-nanometer should be comparable to TSMC 7-nanometer
In terms of chip area, I expect Intel's 10-nanometer technology to offer roughly the same density as TSMC's 7-nanometer technology, based on public statements from both companies. Density is important in the server market, since denser manufacturing technologies allow chipmakers to cram more cores and features into a given chip than would have been possible on a coarser technology.
The power efficiency benefits that come in moving from one technology to a newer one also help, as power efficiency is a key consideration for server chip buyers.
This means that to ensure it doesn't fall behind in terms of server and data-center chip manufacturing technology, and thus product competitiveness, Intel is going to need to get server processors based on its 10-nanometer technology out into the marketplace as quickly as possible.
Intel's second-generation 14-nanometer server processor families, known as Skylake-EP/EN/EX, are expected to arrive in the first half of 2017. DigiTimes recently claimed they'll arrive in the first quarter. Ideally, Intel should release Cannon Lake-EP/EN/EX, the 10-nanometer follow-on products to those processors, about a year later, in the first quarter of 2018.
Such a timeline would very ne aggressive, though, in light of the time it took Intel to transition from one manufacturing technology to the next over the past few generations.
First Intel server product release
It took Intel three and a half years to move its server processor lineup from its 32-nanometer technology to its 22-nanometer technology. It then took the company two and a half years to move that lineup from its 22-nanometer technology to its 14-nanometer technology.
History would suggest that it'll be sometime between the third quarter of 2018 and the third quarter of 2019 before Intel will begin fielding its first server processors built on its 10-nanometer technology. A third-quarter 10-nanometer server chip release, though less than ideal, wouldn't be the end of the world for Intel, but any later would just be plain risky. Ideally, though, Intel ought to get its 10-nanometer server processor lineup out comfortably in the first half of 2018.
What will affect the timing of Intel's 10-nanometer server chip release?
Two major factors could affect Intel's 10-nanometer server processor launch timing. First is simply product readiness. The chip designs need to be finished soon to support product sampling in the second half of 2017 and product launch in the first half of 2018. If Intel's development teams just don't get the chip design done in time, then the schedule must be pushed out.
Even if the design is ready in time for an early 2018 launch, another factor that could influence launch timing will be the health of the 10-nanometer manufacturing process the chips will be built on. Intel's execution with its 14-nanometer process technology was poor, and the schedule for first 10-nanometer product production has already been pushed from late 2015 to the second half of 2017.
If 10-nanometer yields aren't in good shape by the end of 2017, then even with the design ready, Intel will face a difficult choice: Does it ramp up the 10-nanometer server parts and suffer from a suboptimal cost structure, or does it hold off on releasing the new parts until the 10-nanometer technology improves?
If Intel's competitors are as aggressive in trying to bring products built in TSMC's 7-nanometer to this market as TSMC claims, then taking a cost structure hit to maintain product performance and power leadership would be preferable to risking the loss of significant market share and average selling price erosion.
A secret billion-dollar stock opportunity
The world's biggest tech company forgot to show you something, but a few Wall Street analysts and the Fool didn't miss a beat: There's a small company that's powering their brand-new gadgets and the coming revolution in technology. And we think its stock price has nearly unlimited room to run for early in-the-know investors! To be one of them, just click here .
Ashraf Eassa owns shares of Intel. The Motley Fool owns shares of Microsoft. The Motley Fool recommends Intel. Try any of our Foolish newsletter services free for 30 days . We Fools may not all hold the same opinions, but we all believe that considering a diverse range of insights makes us better investors. The Motley Fool has a disclosure policy .
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.