An examination of Intel’s 7nm timeline delay claims

Analysis: How to set up for a worse fall

Intel LogoIn the flaming wreckage of the Q2/2020 Intel analyst call, did the company really come clean on 7nm? SemiAccurate thinks they didn’t, they just set the stage for the future delays and worse problems.

In the Q2/2020 analyst call here, Intel announced that the 7nm process would be delayed by 12 months with the first devices coming out 6 months after initially planned. The first 7nm chips, client not datacenter now, are officially set for late 2022 or early 2023. Technically this is within 12 months or so from the Q4/21 date that Bob Swan gave on the Q1/2020 analyst call but far more than that from the early 2021 dates they have been promising for the past year or more. Is this 12 months of delay? Does Intel believe those dates internally?

Then we come to the question of how a process can be delayed by 12 months and the products only slip by 6 months. This was officially explained by a, “buffer in the planning process” which sounds really good, doesn’t it? If you understand what this really means, things aren’t so rosy. When a new product comes out, especially one on a new process, initial yields are low and wafer starts ramp slowly so as to not waste money as the learning curve is climbed. This is quite normal.

So a new chip might come out on day 1 but the numbers will be really small, not large enough for any OEM to launch a real product line from, but enough for Intel to claim a launch and shipments for revenue. Remember what Intel did when they announced 10nm shipments for revenue? Go back and check the wording on that and compare it to what really shipped and when.

Back to 7nm, these early parts that work are warehoused and once there is both sufficient volume to meet initial demand and sufficiently high wafer throughput to keep demand sated, you launch the product officially. OEMs can make and sell from day 1 and shortages are not an issue. Any guesses how long this ramp and stockpile process usually takes on a new process? Hint: About 6 months historically.

So what Intel is saying with this buffer line that sounds really good on the surface is meant to be something like, “we are really good at planning and minimize any delays through good practice”. What they are actually saying is, “We are screwed and are going to paper launch 7nm before it is ready and before we can provide volume to meet public statements we made earlier. Don’t expect real volumes for 6 months or more after the launch.”. That one doesn’t sound so good now does it? Now do you understand the wording used, and how they got their magic buffer? And does that clue you in to what is happening here too?

But now we get to the multi-billion dollar question, did Intel actually come clean on the call today or is it just a setup for future problems? Is today’s ‘contingency plans’ actually a contingency or are they set and are nothing more than plausible deniability so they can say, “We told you this might happen” on a future call? And does Intel’s internal messaging on 7nm delays match their external messaging on 7nm delays? SemiAccurate looks at some relevant data points.

Note: The following is analysis for professional level subscribers only.

Disclosures: Charlie Demerjian and Stone Arch Networking Services, Inc. have no consulting relationships, investment relationships, or hold any investment positions with any of the companies mentioned in this report.

The following two tabs change content below.

Charlie Demerjian

Roving engine of chaos and snide remarks at SemiAccurate
Charlie Demerjian is the founder of Stone Arch Networking Services and is a technology news site; addressing hardware design, software selection, customization, securing and maintenance, with over one million views per month. He is a technologist and analyst specializing in semiconductors, system and network architecture. As head writer of, he regularly advises writers, analysts, and industry executives on technical matters and long lead industry trends. Charlie is also available through Guidepoint and Mosaic. FullyAccurate