Intel’s 7nm Process Delayed 12 Months, First Product Could Launch In 2023
Use Of External Foundry Is Possible After Latest Delay
It will be another year before we will see 7 nm Intel processors, with late 2022/early 2023 now the target for Intel’s first product on that node. This news comes after Intel’s earnings call yesterday, in which CEO Bob Swan revealed the 12 month production delay, citing a “defect mode in our 7nm process that resulted in yield degradation”.
“We are seeing an approximate six-month shift in our 7nm-based CPU product timing relative to prior expectations. The primary driver is the yield of our 7nm process, which based on recent data, is now trending approximately twelve months behind our internal target.
We have identified a defect mode in our 7nm process that resulted in yield degradation. We’ve root-caused the issue and believe there are no fundamental roadblocks, but we have also invested in contingency plans to hedge against further schedule uncertainty. We’re mitigating the impact of the process delay on our product schedules by leveraging improvements in design methodology such as die disaggregation and advanced packaging.
We have learned from the challenges in our 10nm transition and have a milestone-driven approach to ensure our product competitiveness is not impacted by our process technology roadmap.”
Intel’s struggles with their 10 nm process have been covered extensively over the past couple of years. Cannon Lake barely happened, and, while 10 nm Ice Lake is available in select notebooks, Intel’s highest-performance mobile parts in the 10th Gen family are still built on 14 nm technology.
Perhaps more concerning, and of great interest to followers of the enthusiast PC space, is the fact that to date there have been no Intel desktop processors beyond 14 nm. Every Intel desktop processor since 2015 has been based on the same architecture, with higher core counts and boost clocks following increased competition from AMD, and leveraged by refinements to the venerable 14 nm process.
Skylake lives on, and on… (Image via Wikipedia)
10th Gen Comet Lake is still a Skylake derivative, and has seen limited availability since launching in May. And even if we had 10 nm desktop representation from Intel today, based on what we have seen on the mobile front, Intel would likely need to retain 14 nm CPUs at the top of their product stack to remain competitive with AMD’s current Zen 2 processors.
But Intel is not skipping 10 nm on desktop (as some reports have suggested), after all, with the first Alder Lake parts coming at some point in 2H 2021.
This does not mean that 14 nm is about to be retired; far from it. 10 nm desktop parts might be coming, but Intel’s stated focus “on maintaining an annual cadence of significant product improvements independent of our process roadmap” (emphasis added) is suggestive. We could very well see 14 nm, alongside 10 nm, until the 7 nm transition in 2023.
But perhaps the biggest takeaway from the earnings call was Intel’s apparent willingness to leverage external foundries. The delays have taken their toll, and Intel might have to take a step back from their sole-sourcing stance for processors that began with the i386 in 1985.
“We will continue to invest in our future process technology roadmap, but we will be pragmatic and objective in deploying the process technology that delivers the most predictability and performance for our customers, whether that be our process, external foundry process or a combination of both.”
Intel refused to admit to the public the problems they were having with 10nm by continuing to move the IGP to the CPU package–needlessly costing themselves production capacity and decreasing yields. Their push to improve IGP performance shows a complete lack of understanding of their primary market–business PCs and laptops. Graphics performance above that necessary to do some spreadsheets, presentations, etc. is a complete waste of resources. But they doubled down and continued to put more and more of their limited fab resources into dead die space.
To make things worse, they continued moving forward on their procedure of using a trailing node process to produce chipsets and other non-processor products by moving those to 14nm as well. This would have been the easiest portion of their portfolio to fab externally, but they chose to keep it in house and to eat into their leading edge fab capacity.
I’m glad Jim Keller got out of there before he got caught in the implosion. It also explains why Apple finally chose to pull the cord on their ARM backup parachute. They’ve always used it as a way to threaten Intel that they could always take their business and leave as they weren’t wed to x86 like all of Intel’s other customers. It worked very well fror them. So, no one was surprised they finally did it. But, now we know that it wasn’t over failed negotiations. It was that Intel simply didn’t have any useful products to seel them anymore.
Intel is indeed a mess and its most likely even worse than the earnings call made it sound if you look at whats be happening with Intel the last 4-5 years and read between the lines in this release and other recent ones. Intel needs a serous restructure and that should probably include them spinning off their fabs line AMD did.
Honestly though I think this is good for in the industry. Intel needs to knocked down several pegs from their borderline monopoly position…. Just look at Intel’s product stack in comparison to their competition and their current market position and earnings. No company with Intel’s level product competence should hold the place they do in the market. Its clearly not healthy industry but hopefully and emboldened AMD and ARM can be a change agent.
I find it hilarious that the website suggests a review of the “Intel Lynnfield Core i7-870 and Core i5-750 Processor” from September 8, 2009 and “Intel Pentium D 820 2.8 GHz Dual Core Review” from June 16, 2005 as being “related posts”. Ouch, website, nice burn.