Intel to lay off 15,000 employees(techcrunch.com)
techcrunch.com
Intel to lay off 15,000 employees
https://techcrunch.com/2024/08/01/intel-to-lay-off-15000-employees/
29 comments
He's been in the job roughly 3.5 years. A turnaround that involves building multibillion dollar fab facilities takes longer than that. Many of the problems he is currently dealing with are still leftovers from the prior decade+ of mismanagement. I judge him on two things:
- Whether they execute successfully on the five nodes in four years
- Whether foundry is the right strategy and they successfully lure high profile customers with the opportunity to grow foundry into something that rivals TSMC and Samsung
We will know the answers to both of these by next year. But change horses now when neither question is answered yet? No, I wouldn't do that. Especially after the prior decade of constant CEO turbulence.
- Whether they execute successfully on the five nodes in four years
- Whether foundry is the right strategy and they successfully lure high profile customers with the opportunity to grow foundry into something that rivals TSMC and Samsung
We will know the answers to both of these by next year. But change horses now when neither question is answered yet? No, I wouldn't do that. Especially after the prior decade of constant CEO turbulence.
> He's been in the job roughly 3.5 years. A turnaround that involves building multibillion dollar fab facilities takes longer than that.
Yet mission accomplished in the first 11 months when they paid him 150mil in bonuses.
Yet mission accomplished in the first 11 months when they paid him 150mil in bonuses.
Don’t forget when Jim Keller was senior VP (2018-2020) he was tasked with diagnosing the Intel’s rut and shipwrighting its future. His solution was to sell the fabs as he proclaimed them an albatross, for this he was forced out.
Gelsinger came in and with clownish bravado establishing Intel would pivot to as a third party fab service, choosing the path of maximum pain.
We know the answers now: the global delays, defective product generations. The problems Gelsinger faces are the rearing wall of hubris.
Gelsinger came in and with clownish bravado establishing Intel would pivot to as a third party fab service, choosing the path of maximum pain.
We know the answers now: the global delays, defective product generations. The problems Gelsinger faces are the rearing wall of hubris.
I’m not sure destroying the only significant potential advantage Intel had/(might still have) over its competitors would have been the smartest thing even in hindsight.
The problem is Intel will be judged against the speed of SMIC and other companies, which grows much faster. I understand it’s a completely different game in China, and they have a bigger government backing, but unfortunately 3.5 years is a very long timeline nowadays in today’s competition.
Glesinger is laying off the sales and R&D force he envisioned.
To your questions:
-Are they on track? From everything I've read they're behind and the clock is running out quickly. Intel can't even build a stable product right now and they're about to try and sweep it under the rug. But the concern is meeting some hand-wavy marketing goal?
- The US needs US foundry because of national security interests. That's why the money is there. If foundry isn't/wasn't a strategy Intel has pursued then we're way too far down this path than compared to the node unicorn.
We already know the answers to these questions, because Intel has fallen flat on its face for the last 3 years. I'll give Gelsinger some level of time to make changes, but he didn't put Intel on the right path after his first year: which is more than enough time. And if it's not, then Intel is a failure not worth chasing because it's all rot at the core.
Look at what Gelsinger did at VMware. Nothing other than create, as Sir Topham Hatt says: "Confusion and delay!". They brought about and chased dumpster fire product lines for years and Gelsinger put them on the path to stagnate and get swept up into Broadcom's financial shell game.
You're giving Gelsinger a lot of credit for making zero notable improvements in Intel. And they are where they are right now because of the choices Gelsinger made while he's been in the seat. I'm sorry but sunk cost fallacy isn't a great strategy for Intel right now.
To your questions:
-Are they on track? From everything I've read they're behind and the clock is running out quickly. Intel can't even build a stable product right now and they're about to try and sweep it under the rug. But the concern is meeting some hand-wavy marketing goal?
- The US needs US foundry because of national security interests. That's why the money is there. If foundry isn't/wasn't a strategy Intel has pursued then we're way too far down this path than compared to the node unicorn.
We already know the answers to these questions, because Intel has fallen flat on its face for the last 3 years. I'll give Gelsinger some level of time to make changes, but he didn't put Intel on the right path after his first year: which is more than enough time. And if it's not, then Intel is a failure not worth chasing because it's all rot at the core.
Look at what Gelsinger did at VMware. Nothing other than create, as Sir Topham Hatt says: "Confusion and delay!". They brought about and chased dumpster fire product lines for years and Gelsinger put them on the path to stagnate and get swept up into Broadcom's financial shell game.
You're giving Gelsinger a lot of credit for making zero notable improvements in Intel. And they are where they are right now because of the choices Gelsinger made while he's been in the seat. I'm sorry but sunk cost fallacy isn't a great strategy for Intel right now.
>Intel has fallen flat on its face for the last 3 years
Near as I can tell, Intel has been falling flat on its face since the 10nm debacle started in 2015. They never recovered from that and it's been one problem after another ever since.
Near as I can tell, Intel has been falling flat on its face since the 10nm debacle started in 2015. They never recovered from that and it's been one problem after another ever since.
Intel is a serial fall on it's facer since the 90s. It's just one huge blunder after another. The cash cow keeping them afloat wasn't even their idea they wanted to bet it all on Itanium. They bought a strong mobile SoC starter and poured billions into it starting a decade before iPhone/Android and unceremoniously pissed it a way just as those were exploding, so they could concentrate on Atom. Then they have been strategically sitting around with their thumbs up their asses on GPU/GPGPU for decades.
> Remember that the US taxpayer has sunk over $8.5B in secured funding into Intel already and they are eligible for another $11B in USG loans
This is purely so the US can secure a strategic resource ahead of the East China unification wars in a few years, lets not pretend it has anything to do with retaining jobs.
This is purely so the US can secure a strategic resource ahead of the East China unification wars in a few years, lets not pretend it has anything to do with retaining jobs.
Discussions
(81 points, 7 hours ago, 58 comments) https://news.ycombinator.com/item?id=41133084
(66 points, 7 hours ago, 16 comments) https://news.ycombinator.com/item?id=41133133
(81 points, 7 hours ago, 58 comments) https://news.ycombinator.com/item?id=41133084
(66 points, 7 hours ago, 16 comments) https://news.ycombinator.com/item?id=41133133
Offtopic questions:
> Intel will broadly offer applications for a “voluntary departure” program next week to employees at the company, according to the memo. The company is also announcing a companywide enhanced retirement offering for eligible employees.
What does it mean for anyone to voluntarily depart and what is this enhanced retirement offering?
Never heard of such practices before, hence my mind is very curious.
> Intel will broadly offer applications for a “voluntary departure” program next week to employees at the company, according to the memo. The company is also announcing a companywide enhanced retirement offering for eligible employees.
What does it mean for anyone to voluntarily depart and what is this enhanced retirement offering?
Never heard of such practices before, hence my mind is very curious.
it was 10k the other day, now its 15k. :(
are these all in the bay area?
are these all in the bay area?
Probably plenty in Oregon. A lot of non-core software engineering happens there.
Blaming missing the AI boom on your revenues just shouldn't fly as an excuse.
Good hardware makers should be innovating in other ways not just focusing / relying purely on AI and other tech trends for their revenue.
Does Intel have any chance of getting market share from Nvidia with Gaudi?
PyTorch, TensorFlow, etc can all run reasonably well on Gaudi.
They should have (IMHO Intel was even better positioned than AMD to do that since they generally are much better at software). Considering the obscene margins Nvidia has Intel could still make a very decent profit by selling 1/2 as fast chips for 1/4th the price while consuming 50% more power.
Realistically seeing how mismanaged to company the chance is probably pretty low though..
Realistically seeing how mismanaged to company the chance is probably pretty low though..
Gaudi was made by an Israeli company Intel acquired in 2019 (not an internal project).
Gaudi 3 has PCIe 4.0 (vs. H100 PCIe 5.0, so 2x the bandwidth). It's strange for Intel, of all vendors, to lag behind in PCIe. And Nvidia has SXM for larger models (5x bandwidth).
"N5, PCIe 4.0, and HBM2e. This chip was probably delayed two years." -wmf
Gaudi 3 has PCIe 4.0 (vs. H100 PCIe 5.0, so 2x the bandwidth). It's strange for Intel, of all vendors, to lag behind in PCIe. And Nvidia has SXM for larger models (5x bandwidth).
"N5, PCIe 4.0, and HBM2e. This chip was probably delayed two years." -wmf
not only did they miss smartphones and now AI, they're getting whooped by AMD in servers and are close to irrelevant in low power / laptops.
AMD also missed smartphones, and is barely a player in AI.
It's salvageable. I think people forget just how bad Bulldozer and friends were. AMD was on its death bed, and put aces in their places and turned it around.
Intel can do it just the same, but I'm beginning to feel Gelsinger isn't the person for the job.
It's salvageable. I think people forget just how bad Bulldozer and friends were. AMD was on its death bed, and put aces in their places and turned it around.
Intel can do it just the same, but I'm beginning to feel Gelsinger isn't the person for the job.
As you said AMD wasn’t even in a position to do anything about mobile.
Intel on the other hand had the best high-end ARM chips in the early 2000s. XScale would have been the default choice for the iPhone and all the other smartphones. Had they continued developing it it’s unlikely anyone could have caught up that easily. They were better positioned to dominate the ARM market than Qualcomm (and possibly ARM itself).
Then due to hubris/delusion/? they decided to scrap XScale and develop low-power x86 chips and everyone knows how that turned out..
Intel on the other hand had the best high-end ARM chips in the early 2000s. XScale would have been the default choice for the iPhone and all the other smartphones. Had they continued developing it it’s unlikely anyone could have caught up that easily. They were better positioned to dominate the ARM market than Qualcomm (and possibly ARM itself).
Then due to hubris/delusion/? they decided to scrap XScale and develop low-power x86 chips and everyone knows how that turned out..
> are close to irrelevant in low power / laptops.
Last gen chips (made at TSMC) seem to be pretty decent? Not 100% sure but even Meteor Lake seems to be more on less on par (or pretty close) with Qualcomm CPU speed/power usage wise while having significantly better GPUs. Lunar Lake might also be a significant improvement.
Last gen chips (made at TSMC) seem to be pretty decent? Not 100% sure but even Meteor Lake seems to be more on less on par (or pretty close) with Qualcomm CPU speed/power usage wise while having significantly better GPUs. Lunar Lake might also be a significant improvement.
And ARM is eating their lunch in cloud, too.
[dupe]
More discussion on official release: https://news.ycombinator.com/item?id=41133084
More discussion on official release: https://news.ycombinator.com/item?id=41133084
I don't know if laying off people is the right move here.
yeah intel is in for a bit of pain as they reconfigure. but they're smart. they just need to execute with the ferocity of a mongol general.
yeah intel is in for a bit of pain as they reconfigure. but they're smart. they just need to execute with the ferocity of a mongol general.
Remember that the US taxpayer has sunk over $8.5B in secured funding into Intel already and they are eligible for another $11B in USG loans [0]. Gelsinger should not be the guy in charge of this anymore. Hopefully the US calls into question the path Intel is on as it is a matter of national security in the long run.
[0] https://www.intel.com/content/www/us/en/newsroom/news/us-chi...