Intel has just published a news release on its website stating that Jim Keller has resigned from the company, effective immediately, due to personal reasons.

Jim Keller was hired by Intel two years ago to the role as Senior Vice President of Intel’s Silicon Engineering Group, after a string of successes at Tesla, AMD, Apple, AMD (again), and PA Semiconductor. As far as we understand, Jim’s goal inside Intel was to streamline a lot of the product development process on the silicon side, as well as providing strategic platforms though which future products can be developed and optimized to market. We also believe that Jim Keller has had a hand in looking at Intel’s manufacturing processes, as well as a number of future products.

Intel’s press release today states that Jim Keller is leaving the position on June 11th due to personal reasons. However, he will remain with the company as a consultant for six months in order to assist with the transition.

As a result of Jim’s departure, Intel has realigned some of its working groups internally with a series of promotions.

  1. Sundari Mitra, the former CEO and founder of Net Speed, will lead a newly created IP Engineering Group.
  2. Gloria Leong will head the Xeon Performance Group
  3. Gene Scuteri will head the Xeon and Networking Engineering Group
  4. Uri Frank and Boyd Phelps will lead the Client Engineering Group
  5. Daaman Hejmadi will lead the Design Enablement Group
  6. Navid Shahriari will continue to lead the Manufacturing and Product Engineering Group

Jim Keller’s history in the industry has been well documented – his work has had a significant effect in a number of areas that have propelled the industry forward. This includes work on Apple’s A4 and A5 processors, AMD’s K8 and Zen high-level designs, as well as Tesla’s custom silicon for self driving, which Tesla’s own competitors have said put the company up to seven years ahead.

With our interview with Jim Keller, several weeks after taking the job at Intel, we learned that Keller went in to the company with a spanner. Keller has repeatedly said that he’s a fixer, more than a visionary, and Intel would allow him to effect change at a larger scale than he had ever done previously.

From our interview:

JK: I like the whole pipeline, like, I've been talking to people about how do our bring up labs and power performance characterization work, such as how does our SoC and integration and verification work? I like examining the whole stack. We're doing an evaluation on how long it takes to get a new design into emulation, what the quality metrics are, so yeah I'm all over the place.

We just had an AI summit where all the leaders for AI were there, we have quite a few projects going on there, I mean Intel's a major player in AI already, like virtually every software stack runs on Xeon and we have quite a few projects going on. There's the advanced development stuff, there's nuts and bolts execution, there's process and methodology bring up. Yeah I have a fairly broad experience in the computer business. I'm a ‘no stone unturned’ technical kind of person – when we were in Haifa and I was bugging an engineer about the cleanliness of the fixture where the surface mount packages plug into the test boards.

Jim’s history has shown that he likes to spend a few years at a company and move on to different sorts of challenges. His two year stint at Intel has been one of his shortest tenures, and even recently Fortune published a deep expose on Jim, stating that ‘Intel is betting its chips on microprocessor mastermind Jim Keller’. So the fact that he is leaving relatively early based on his previous roles is somewhat different.

Intel’s press release on the matter suggests that this has been known about for enough time to rearrange some of the working groups around to cover Jim’s role. Jim will be serving at Intel for at least another six months it seems, in the role of a consultant, so it might be that long before he lands another spot in the industry.

It should be noted that Jim Keller is still listed to give one of the keynote addresses at this year’s Hot Chips conference on behalf on Intel. We will update this story if that changes.

This news item was updated on 17th June with information regarding the new rearrangement. Points 2 and 4 were added, while (the new) 5 was adjusted.

Related Reading


Comments Locked


View All Comments

  • AshlayW - Friday, June 12, 2020 - link

    Let's not forget that Intel's budget is orders of magnitude higher than AMD's. Let's not forget that Skylake doesn't compete with Zen2, not even close, in metrics that matter such as performance per watt, or compute density (cores per socket) / MT performance and I/O (for servers).

    Your comment is spoken like a clueless gamer who thinks that a few extra % average FPS in a latency-sensitive workload on a dead-end architecture without chiplets and any future is the last word in microrprocessor technology. Let's also not forget the security vulnerabilities.

    AMD has innovated more in the processor industry between 2017 and today, than Intel has in the last decade. And you need to accept that.
  • Gondalf - Friday, June 12, 2020 - link

    Between 20217 and now i can see only an Ark----->Zen, Zen 2 is only a small revamping (comparable with Skylake in per core raw performance) and Zen 3 that is is only a better arrangement of the cache stack of Zen 2.
    More or less have done Intel with Sunny Cove (better SMT), Willow cove (even better SMT), and soon Golden Cove ( nice strong surprise for many) and Ocean Cove (another strong news).

    Accept this :)

    The only contribute of AMD is to have joined with TSMC ( so kudos to TSMC, definitively NO to AMD ). The main AMD sin is the chiplet desing without a really good interconnection.
  • schujj07 - Friday, June 12, 2020 - link

    "comparable with Skylake in per core raw performance" only when the Skylake CPU has a 10+% clock speed advantage and using a lot more power. We have seen the i7-7700k vs Ryzen 3 3300X review and the 7700k loses far more often than it wins. Both are 4c/8t CPUs but the 7700k has less than a 10% clock speed advantage. That means that Zen 2's IPC advantage over Skylake allows its lower clocks to overtake the Skylake chip.

    Sunny Cove has a better IPC on paper than Zen 2, however, when running none SPEC applications it seems to falter. Look at the Acer Swift 3 review that has the best Ice Lake i7 and the near best Ryzen 4700U. Yes the 4700U has double the cores and a 100MHz boost clock speed advantage. However, Ice Lake is supposed to have a 7ish% IPC advantage over Zen 2 but loses on many ST applications by more than the 2.5% boost clock disadvantage should allow.
  • iranterres - Friday, June 12, 2020 - link

    I just read jibberish about it chiplet and node and blah blah blah here. What you all need to accept are just two simple facts:

    - Intel's comfortable market position in the past decade made them seek architecture evolution, not design evolution. Now their game is milked up.
    - AMD went through the last decade with a fundamentaly flawed design (bulldozer/vishera) that turned up also as an architecture dead-end.

    Now it is just the same history, with sides exchanged.

    And yes, fanboys will be fanboys, both sides. You suck.
  • AshlayW - Friday, June 12, 2020 - link

    Don't bring politics into the tech discussion please. Because the "Israel is running a Gulag" thing is subjective and political in nature.

    Intel's technology inferiority, however, is objective.
  • willis936 - Friday, June 12, 2020 - link

    It’s impossible to separate them. It would be nice and clean if you could, but that isn’t reality.
  • Oxford Guy - Thursday, June 11, 2020 - link

    Effective immediately = six months from now.
  • eastcoast_pete - Thursday, June 11, 2020 - link

    Most people of Jim Keller's seniority and stature have significant non-compete clauses in their contracts, which often explains why they remain "consultants" for up to and over a year after they either resign or get fired. Basically, one has to "age out" of the inside knowledge before moving on, but that time is paid for.
  • MooseNSquirrel - Friday, June 12, 2020 - link

    Non compete is basically non enforceable in California.
  • rrinker - Friday, June 12, 2020 - link

    Just because they'll lose in the end doesn't mean a company with deep pockets and lots of corporate lawyers can't make your life a living hell. Non-competes are not enforceable in many places, especially since with specialized jobs it amounts to slave labor if it could be enforced that you can essentially never work in your industry again.

Log in

Don't have an account? Sign up now