4 ms·

Intel to lay off 15,000 employees

86 points by bulla 2y ago

bimguy 2y ago
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.
gnabgib 2y ago
Discussions (81 points, 7 hours ago, 58 comments) https://news.ycombinator.com/item?id=41133084 https://news.ycombinator.com/item?id=41133084 (66 points, 7 hours ago, 16 comments) https://news.ycombinator.com/item?id=41133133 https://news.ycombinator.com/item?id=41133133
cranberryturkey 2y ago
it was 10k the other day, now its 15k. :( are these all in the bay area?
- dcdc123 2y ago
  Probably plenty in Oregon. A lot of non-core software engineering happens there.
windexh8er 2y ago
Gelsinger needs to resign / be fired by the board. He is not the right person leading Intel, he wasn't the right person leading VMware. He's a CEO with a superiority complex. 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-chips-act-intel-direct-funding.html https://www.intel.com/content/www/us/en/newsroom/news/us-chi...
- WoodenChair 2y ago
  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.
  - kredd 2y ago
    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.
  - windexh8er 2y ago
    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.
laweijfmvo 2y ago
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.
- silisili 2y ago
  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.
  - ffgjgf1 2y ago
    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..
- ffgjgf1 2y ago
  > 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.
- alecco 2y ago
  And ARM is eating their lunch in cloud, too.
jamesblonde 2y ago
Does Intel have any chance of getting market share from Nvidia with Gaudi? PyTorch, TensorFlow, etc can all run reasonably well on Gaudi.
- ffgjgf1 2y ago
  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..
- alecco 2y ago
  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
ChrisArchitect 2y ago
[dupe] More discussion on official release: https://news.ycombinator.com/item?id=41133084 https://news.ycombinator.com/item?id=41133084
n_ary 2y ago
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.
- adam_beck 2y ago
  They are just severance packages.
  - zeusk 2y ago
    No, they aren't. Intel is old enough to have an actual "retirement" plan and not just a 401k match program.
dzonga 2y ago
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.
Havoc 2y ago
I've seen 20k, 15k and 10k mentioned. Seems like nobody knows what's going on
- rsanek 2y ago
  when I search Intel layoff on Google, the only number I'm seeing is 15k

86 points by bulla 2y ago

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.

gnabgib 2y ago

Discussions (81 points, 7 hours ago, 58 comments) https://news.ycombinator.com/item?id=41133084 https://news.ycombinator.com/item?id=41133084 (66 points, 7 hours ago, 16 comments) https://news.ycombinator.com/item?id=41133133 https://news.ycombinator.com/item?id=41133133

cranberryturkey 2y ago

it was 10k the other day, now its 15k. :( are these all in the bay area?

dcdc123 2y ago
Probably plenty in Oregon. A lot of non-core software engineering happens there.

windexh8er 2y ago

Gelsinger needs to resign / be fired by the board. He is not the right person leading Intel, he wasn't the right person leading VMware. He's a CEO with a superiority complex. 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-chips-act-intel-direct-funding.html https://www.intel.com/content/www/us/en/newsroom/news/us-chi...

WoodenChair 2y ago
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.
- kredd 2y ago
  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.
- windexh8er 2y ago
  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.

laweijfmvo 2y ago

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.

silisili 2y ago
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.
- ffgjgf1 2y ago
  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..
ffgjgf1 2y ago
> 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.
alecco 2y ago
And ARM is eating their lunch in cloud, too.

jamesblonde 2y ago

Does Intel have any chance of getting market share from Nvidia with Gaudi? PyTorch, TensorFlow, etc can all run reasonably well on Gaudi.

ffgjgf1 2y ago
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..
alecco 2y ago
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

ChrisArchitect 2y ago

[dupe] More discussion on official release: https://news.ycombinator.com/item?id=41133084 https://news.ycombinator.com/item?id=41133084

n_ary 2y ago

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.

adam_beck 2y ago
They are just severance packages.
- zeusk 2y ago
  No, they aren't. Intel is old enough to have an actual "retirement" plan and not just a 401k match program.

dzonga 2y ago

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.

Havoc 2y ago

I've seen 20k, 15k and 10k mentioned. Seems like nobody knows what's going on

rsanek 2y ago
when I search Intel layoff on Google, the only number I'm seeing is 15k