Intel's 10nm Is Broken, Delayed Until 2019

Intel announced its financial results today, and although it posted yet another record quarter, the company unveiled serious production problems with its 10nm process. As a result, Intel announced that it is shipping yet more 14nm iterations this year. They'll come as Whiskey Lake processors destined for the desktop and Cascade Lake Xeons for the data center. 

Intel nailed a nearly perfect quarter with top-line numbers that include record Q1 revenue of $16.8 billion, which is up 16% year-over-year (YoY). The company also raised its guidance for the full year to $67.5 billion, which is a $2.5 billion bump over its previous guidance, but we'll circle back to those numbers later.

The 10nm Problems

Overall, Intel had a stellar quarter, but it originally promised that it would deliver the 10nm process back in 2015. After several delays, the company assured that it would deliver 10nm processors to market in 2017. That was further refined to the second half of this year.

On the earnings call today, Intel announced that it had delayed high-volume 10nm production to an unspecified time in 2019. Meanwhile, its competitors, like TSMC, are beginning high volume manufacturing of 7nm alternatives.

Recent semiconductor node naming conventions aren't based on traditional measurements, so they're more of a marketing exercise than a science-based metric. That means that TSMC's 7nm isn't entirely on par with Intel's 10nm process. However, continued process node shrinks at other fabs show that other companies are successfully outmaneuvering the production challenges of smaller lithographies.

Intel's CEO Brian Krzanich repeatedly pressed the point that the company is shipping Cannon Lake in low volume, but the company hasn't pointed to specific customers or products. And we've asked. As we pointed out earlier this year, the delay may seem a minor matter, but Intel has sold processors based on the underlying Skylake microarchitecture since 2015, and it's been stuck at the 14nm process since 2014. That means Intel is on the fourth (or fifth) iteration of the same process, which has hampered its ability to bring new microarchitectures to market. That doesn't bode well for a company that regularly claims its process node technology is three years ahead of its competitors.

Krzanich explained that the company "bit off a little too much on this thing" by increasing 10nm density 2.7X over the 14nm node. By comparison, Intel increased density by only 2.4X when it moved to 14nm. Although the difference may be small, Krzanich pointed out that the industry average for density improvements is only 1.5-2X per node transition. Because of the production difficulties with 10nm, Intel has revised its density target back to 2.4X for the transition to the 7nm node. Intel will also lean more on heterogeneous architectures with its EMIB technology (which we covered here).

10nm is Intel's last process based on traditional photolithography, and though Krzanich didn’t dive deep into details, he listed the lithography technique as a significant contributor to the low 10nm yields. The company will switch to EUV at 7nm. Currently, Intel's multipatterning process is generating too many yield-reducing defects to produce 10nm cost-effectively. Krzanich says the company has identified the issue and is moving to correct it, but the fixes will take an unspecified amount of time to impact yields significantly. Intel was unwilling to commit to high volume production in the first half of 2019, so it's possible 10nm will be delayed until the second half of the year. 

We continue to make progress on our 10-nanometer process. We are shipping in low volume and yields are improving, but the rate of improvement is slower than we anticipated. As a result, volume production is moving from the second half of 2018 into 2019. We understand the yield issues and have defined improvements for them, but they will take time to implement and qualify. We have leadership products on the roadmap that continue to take advantage of 14-nanometer, with Whiskey Lake for clients and Cascade Lake for the data center coming later this year.

Krzanich’s statement says that while Intel has defined the fixes, it hasn’t actually tested them in high-volume manufacturing. That means the company could turn back to the drawing board soon if the fixes aren’t effective.

That's likely the impetus for Intel's confirmation today that it realigned its critical Technology and Manufacturing Group, which produces the company's silicon, under new leadership. Intel also made a timely pre-earnings-call announcement today that it had brought in famed chip architect Jim Keller to head up its silicon design initiatives. This was likely a move to assure investors that the 10nm production issues have Intel's full attention.

Unfortunately, process technologies require extensive incubation periods, so it could take some time before leadership changes have a significant impact on Intel's roadmap. Intel's obviously bringing the pieces together quickly, but its competitors, such as AMD, are executing well on their future architectures. AMD already has working 7nm GPUs in its labs and projects it will sample 7nm EPYC 2 processors this year. Both will be in volume production early next year.

Thoughts

Intel's late 10nm process has led to stagnation on the microarchitecture development front, and the problems are even larger than they appear on the surface. As the financial results below clearly outline, the company is successfully diversifying into AI, the data center, autonomous driving, 5G, FPGAs, and IoT, among other climes. It's even added GPUs to the list.

Unfortunately, Intel's process technology touches every segment of all that tech, as well as the chips that power them. 10nm's late arrival could hamper Intel's competitiveness in nearly all of those segments, and all this comes as Intel is expanding into new segments that already have dominant and entrenched players with deep pockets. Krzanich did point out that Intel has improved 14nm's performance by 70% since its debut in 2014, but the company will surely reach a diminishing point of returns soon.

Intel's relatively flat R&D spend (+3%) is certainly not encouraging, given the current climate. The company will likely switch to second-gen 10nm+ due to its yield issues with 10nm, but it did not confirm the change during the earnings call. Krzanich did say the company will not skip to the 7nm node. Instead, it will apply its learning from the 10nm node to the 7nm process.

Krzanich also admitted that the company's density lead over competing fabs is shrinking. Intel has long been the keeper of the Moore's Law flame, and the company has continued to insist that the Law is still alive long after other companies have conceded that it expired. We'll have to see if Intel changes its messaging, but we're a long time removed from the Tick-Tock cadence. Considering that Intel hasn't delivered a smaller process in significant volumes since 2014, it's fair to say that the original Moore's Law is officially dead.

The Financials

Intel's Client Computing Group (CCG) posted strong financial results. The group focuses on processors for laptops and desktops and has been under an extended assault from AMD, which continues to enjoy brisk momentum. Intel's CCG group posted $8.2 billion in revenue, a 3% YoY gain, but this is largely due to increased average selling prices and strength in enthusiast processor sales.

The company suffered an 8% decline in notebook processor sales and a 6% reduction in desktop CPU sales compared to last year. Intel cites a shrinking addressable market as a key contributor, but a recent market report from IDC says that PC volumes were flat, while competing bean counter Gartner says that PC shipments declined 1.4% compared to the first quarter of last year. Intel did notch a 4% gain in laptop sales compared to last year.

It isn't clear how much of this decline is due to AMD's continued success, but AMD told us that it has gained up to 50% of the CPU sales to DIYers at retailers like Amazon and Newegg. AMD also reported a 95% increase in its CPU and graphics business last quarter, but it's hard to determine how much of that stems from processor sales alone.

Intel's Data Center Group (DCG) grew revenue 24% on sales of pricey Xeon Scalable models, but Intel is reducing its guidance for server sales in the second half of the year, citing "tougher competition going into the second half." That is likely in reference to the combination of AMD's EPYC processors and new ARM competitors from Qualcomm.

Intel has spent the last year transitioning to "data-centric" businesses, which are largely composed of data center processors. This comes as the company reduces its reliance on the bread-and-butter PC segment. DCG contributed 46% of Intel's revenue in Q1, so that plan is apparently on track. 

Intel's Non-Volatile Storage Group segment produces NAND and 3D XPoint. NSG revenue is also up 20%, while the Programmable Solutions Group, which produces FPGAs, is up 17%.

Create a new thread in the News comments forum about this subject
This thread is closed for comments
37 comments
Comment from the forums
    Your comment
    Top Comments
  • rhysiam
    So let me get this straight: competition from AMD and issues with the 10nm process meant that Intel, fuelled by a caffeine hit, rushed "Coffee Lake" to market. This was to remain competitive until they could come out all-guns-blazing with the true successor, "Cannon Lake". Now instead, as 10nm problems persist, we'll all have to drown our sorrows with "Whisky Lake".

    Is that about the measure of it?
  • Other Comments
  • InvalidError
    Yet another 14nm refinement year, yay!

    Intel used to be a solid two years ahead of everyone else, that lead shrunk to a year on 16/14nm, now it looks like Intel might be falling a year behind on 10/7nm.

    This will hopefully have interesting consequences on competitiveness and pricing pressure.
  • dudmont
    70% performance improvement in the 14nm node? How much of that is entirely from adding cores?
    Intel used to be worth 8% IPC improvement to each new series. It looks to me like AMD is about 1 generation away from true parity and they still have em beat in cores and threads by a good amount. Is all the flurry in hiring a sign of someone hitting the panic button?
    Good luck in Houston :). Here's hoping my Vikes get their G.
  • modeonoff
    Does that mean CPUs such as Cascade Lake that should have the Meltdown and Spectre issues removed at hardware level will be delay to 2019?