News Intel lays off hundreds of engineers in California, including chip design engineers and architects — automotive chip division also gets the axe

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Are you sure? Since when? I'm sure I heard about at least 3 board members leaving, late last year.
Unless I'm remembering wrong they added two technical board members (both have semi experience) last year right after punting Gelsinger. This was the big change that needed to happen period, but they were still very outnumbered. Before that the board was seven finance and five technical two of which were semi (including Gelsinger). Then I think when Lip-Bu Tan was hired as CEO he rejoined the board.

I must have missed the news about the three retirements or my brain is still malfunctioning today. The three leaving (two are longer term board members) are two finance and one technical so that's a five/six split in favor of technical (four semi).
 
I love that the bashing of intel you do is almost blind hatred at this point...
I dont hate intel i just dont kiss there bum for EVERY bad decision the seem to make which has put them so far behind AMD which is forcing AMD to be the next Nvidia ..

Pushing out BS graphs and projection figures means nothing when they are clearly drowning as a company..

They still lack X3d they still lack efficient cpu's and still lack socket life !!

core ultra failed big time for them

14th was a joke over 13th !!

SO by all means call me a hater of intel but anyone with half a brain would understand the logic that we as consumers need healthy competition between the 2 CPU makers and at the moment AMD are crushing them ..

which will lead to over priced stuff from AMD because their competition is drowning in failure !!

Ive seen the rumours of the next gen intel stuff and it may be compelling but will it beat AMD or will it get crushed again ??

If itel offer something that crushes AMD and promises me long like span i will swap to Intel otherwise why would i buy an inferior product !!
 
core ultra failed big time for them

14th was a joke over 13th !!
Yeah, it's not unlike how they got stuck on 14nm Skylake-era technology. Except now, they're stuck on 2022-era Raptor Lake. Or, you could argue 2021-era Alder Lake.

In fact, they're still launching new products based Alder Lake silicon, to this very day - just rebadged as something newer. Check out some of the models in their Bartlett Lake generation! At least, in 14nm "Skylake" era (from 2015 to 2020), most of that silicon was actually new and not just rebadged.

Ive seen the rumours of the next gen intel stuff and it may be compelling but will it beat AMD or will it get crushed again ??
Nova Lake sounds good on paper, but so did Arrow Lake. So, I really don't know. I think a lot of it comes down to execution and not pushing a half-finished product out the door.
 
I used to think Terry was Pat Gelsinger but it doesn’t make sense; I now suspect he might be Lip-Bu Tan.

Damn you... I almost lost a good keyboard.

Regards 😆
Yeah because it takes some kind of genius to understand that if intel wouldn't spend ~30B a year due to the FABs they would be making money, anything lower in intellect than gelsinger/lip just can't understand this concept...it's just way too complicated to grasp.
I dont hate intel i just dont kiss there bum for EVERY bad decision the seem to make which has put them so far behind AMD which is forcing AMD to be the next Nvidia ..

Pushing out BS graphs and projection figures means nothing when they are clearly drowning as a company..

They still lack X3d they still lack efficient cpu's and still lack socket life !!

core ultra failed big time for them

14th was a joke over 13th !!

SO by all means call me a hater of intel but anyone with half a brain would understand the logic that we as consumers need healthy competition between the 2 CPU makers and at the moment AMD are crushing them ..

which will lead to over priced stuff from AMD because their competition is drowning in failure !!

Ive seen the rumours of the next gen intel stuff and it may be compelling but will it beat AMD or will it get crushed again ??

If itel offer something that crushes AMD and promises me long like span i will swap to Intel otherwise why would i buy an inferior product !!
Yup that's what happens when somebody just goes by how they feel about companies, then they can just make anything they want up and there are no numbers to confuse them.
You do you.
 
Yeah because it takes some kind of genius to understand that if intel wouldn't spend ~30B a year due to the FABs they would be making money, anything lower in intellect than gelsinger/lip just can't understand this concept...it's just way too complicated to grasp.
Wall St. is literally full of people who get that, and want Intel to spin out its fabs. That way, it'd be purely fabless, like AMD and Nvidia.

When Intel's fabs are working for them, they do improve margins. However, it requires lots of ongoing investment in not just new physical plant, but also R&D. Striking the right balance has proven very challenging for Intel, lately.
 
Striking the right balance has proven very challenging for Intel, lately.
There hasn't been any balance yet...
Up to today they are just laying the groundwork, as far as I know only the two 18A FABs are ready and those haven't started making any money yet, they might have started production or are about to, but didn't produce any money yet.
They even made their full mainstream lineup at TSMC the last year so not even their older FABs are making them money/all the money they could be making them, at least they do use them for server parts.
Up to this day they have only been paying, which is what makes the analysts/investors/whoever anxious.

It's when the new FABs (even just 18A) will start production where we will see (well intel will see, we will see a good time later) if they can strike the right balance or not, until then it's just a game of perseverance to even get going.
 
  • Like
Reactions: shady28
There hasn't been any balance yet...
Up to today they are just laying the groundwork, as far as I know only the two 18A FABs are ready and those haven't started making any money yet, they might have started production or are about to, but didn't produce any money yet.
But, it's a pipeline. And even if they can reuse the existing buildings for newer nodes, the equipment is getting more expensive and I'm sure there are costs to retrofit an old fab for newer equipment.

You talk a lot about these costs being a one-time thing, but I don't believe they would be (i.e. if Intel were serious about staying in the manufacturing game).

They even made their full mainstream lineup at TSMC the last year so not even their older FABs are making them money/all the money they could be making them, at least they do use them for server parts.
Yeah, and that compounds their problems. However, if I understand correctly, they're also talking about continuing to use TSMC into the foreseeable future. So, it's also not just a one-time issue.
 
But, it's a pipeline. And even if they can reuse the existing buildings for newer nodes, the equipment is getting more expensive and I'm sure there are costs to retrofit an old fab for newer equipment.

You talk a lot about these costs being a one-time thing, but I don't believe they would be (i.e. if Intel were serious about staying in the manufacturing game).
If the long term costs where as high as the up front costs then nobody would be doing it...
Now if intel foundry does zero dollars for years then yes any modernization will be difficult and expensive.
Otherwise the income will by far outweigh the costs.
Yeah, and that compounds their problems. However, if I understand correctly, they're also talking about continuing to use TSMC into the foreseeable future. So, it's also not just a one-time issue.
Sure as long as they will make a margin of 50% ....so they will be allowed to use TSMC but in reality it will never or very rarely happen.
https://www.tomshardware.com/tech-i...ust-deliver-50-percent-to-get-the-green-light
 
If the long term costs where as high as the up front costs then nobody would be doing it...
The way to make it viable is to increase volume, which was the point of IFS. If successful, it would enable them to structure their fab investments a lot more like TSMC does.

Now if intel foundry does zero dollars for years then yes any modernization will be difficult and expensive.
Not just difficult and expensive, they'd fall behind and be unable to claw their way back to competitiveness. As it is, being competitive is just barely within reach!

There are real examples of where this happened: Global Foundries and UMC. They decided to milk their legacy nodes, but now they're getting worried as China looks set to snatch that meal ticket away from them. So, now they're trying to figure out how to modernize and it seems they have few ideas other than merging to pool their resources.
 
  • Like
Reactions: thestryker
Not just difficult and expensive, they'd fall behind and be unable to claw their way back to competitiveness.
Well the good thing is that they only need to stay competitive if they have customers that are willing to pay more for better nodes, and if they have such customers they will have the money to do it.
As far as I know nobody is forced to modernize all their fabs all at once all of the years...you do as much as you need to, as much as there will be volume for.
Modernizing one FAB, that I assume is already build to be modern, for example can't possibly be more than the 30B they spend now per year, and it won't happen every year, unless there is high demand which would also pay for it.
As it is, being competitive is just barely within reach!
Based on what?!
Also which kind of competitive?
Because as long as they make a good enough product that enough people will pay for then that's all they need, even if it's not competitive, as long as it's lucrative.
 
Mongolion flame-thrower doin' some (probably good) work.

Hopefully this new push works out in the end; I suspect that it was either this or a slow death for Intel.
It's still going to be a slow death for Intel. This year the competition from AMD and Qualcomm is going to get even fiercer.

And Intel still hasn't matched the competitors last gen products lol
 
It's still going to be a slow death for Intel. This year the competition from AMD and Qualcomm is going to get even fiercer.

And Intel still hasn't matched the competitors last gen products lol
Lunar Lake is good, but addresses a limited slice of the market - especially when you consider pricing.

Nova Lake should be pretty good. The question is exactly when it will be ready. If it can launch well in advance of Zen 6, that might give it a good runway.

In the meantime, Panther Lake is going to be very interesting to watch. It'll feature new cores and it should tell us most of what we want to know about 18A.

Intel's money-printing days are truly behind it, but I wouldn't count them out just yet. As for their medium/long-term prospects, I think that has a lot to do with how their fab R&D is going for 1.4A and beyond.
 
Nova Lake should be pretty good. The question is exactly when it will be ready. If it can launch well in advance of Zen 6, that might give it a good runway.
I think it's fair to say there's pretty much no chance of beating AMD to market unless they move it forward. There won't be a major desktop SKU this year and Intel typically schedules fall. Best case scenario they're probably coming out around the same time.
 
  • Like
Reactions: bit_user
I think it's fair to say there's pretty much no chance of beating AMD to market unless they move it forward. There won't be a major desktop SKU this year and Intel typically schedules fall. Best case scenario they're probably coming out around the same time.
Wait, didn't it slip from launching at the end of this year to early next? Or am I just confused?
 
Wait, didn't it slip from launching at the end of this year to early next? Or am I just confused?
NVL was always a 2026 part it's PTL that was supposed to be 2025 and basically slipped to 2026 (it might be like LNL where a handful slipped through near "launch"). That's why there was supposed to be an ARL refresh of some sort this year though the plans for that have changed multiple times. The most recent leak regarding that was 3 months ago Jaykihn saying it's only K SKUs. The longer we go without solid information the more I think these will just be binned die and no more if it even happens.

I wouldn't really be surprised if there was nothing new on desktop until NVL. It makes sense to string along ADL/RPL die as long as possible to keep the DUV fabs utilization up and this probably also explains the new BTL die even existing.
 
  • Like
Reactions: bit_user
Yeah, it's not unlike how they got stuck on 14nm Skylake-era technology. Except now, they're stuck on 2022-era Raptor Lake. Or, you could argue 2021-era Alder Lake.

In fact, they're still launching new products based Alder Lake silicon, to this very day - just rebadged as something newer. Check out some of the models in their Bartlett Lake generation! At least, in 14nm "Skylake" era (from 2015 to 2020), most of that silicon was actually new and not just rebadged.


Nova Lake sounds good on paper, but so did Arrow Lake. So, I really don't know. I think a lot of it comes down to execution and not pushing a half-finished product out the door.
Personally I don’t see any win for intel in the foreseeable future more so if their next get is on a new socket ..

If AMD I’m just guessing 10k series is still am5 then AMD have already won that battle because why would I bother with a new mobo when I can most likely use my x670e board ??

Intel need to come out swinging and swinging BIG offer longer socket life span and x3d with efficiency..

Something to scare AMD to put them back in the good graces with consumers as of now they need something very special next gen ..

I’m not sure they can stay alive with another failure..
 
Last edited:
Lunar Lake is good, but addresses a limited slice of the market - especially when you consider pricing.

Nova Lake should be pretty good. The question is exactly when it will be ready. If it can launch well in advance of Zen 6, that might give it a good runway.

In the meantime, Panther Lake is going to be very interesting to watch. It'll feature new cores and it should tell us most of what we want to know about 18A.

Intel's money-printing days are truly behind it, but I wouldn't count them out just yet. As for their medium/long-term prospects, I think that has a lot to do with how their fab R&D is going for 1.4A and beyond.

Just to give you an idea of where Intel sits, in 2024 they made $981M from interest. They had $196B in assets and $91B in liabilities.

If Intel were to pull back R&D to pre-Gelsinger levels, they'd be about zero on earnings. If they shut it down to near zero, they'd make about $5B.

Since 18A hasn't started generating revenue yet, the bet is pretty clear. Also, as far as future costs, Intel is the only one who has already purchased and invested in developing with high-NA lith.

TSMC hasn't done that yet.

I don't know how it will turn out, but that strategy sounds ominously like Intel's pre-Gelsinger one. i.e. "We don't need to spend on EUV, we can do it with DUV on a shoestring R&D budget" 10nm/7nm fiasco. Only this time, it's coming from TSMC.

Worth noting that TSMC is way behind their original N2 timeline. They slipped by a year.

 
Personally I don’t see any win for intel in the foreseeable future more so if their next get is on a new socket ..

If AMD I’m just guessing 10k series is still am5 then AMD have already won that battle because why would I bother with a new mobo when I can most likely use my x670e board ??

Intel need to come out swinging and swinging BIG offer longer socket life span and x3d with efficiency..

Something to scare AMD to put them back in the good graces with consumers as of now they need something very special next gen ..

I’m not sure they can stay alive with another failure..
You're more pessimistic than I am, but I guess we'll just have to wait and see.
 
Since 18A hasn't started generating revenue yet, the bet is pretty clear. Also, as far as future costs, Intel is the only one who has already purchased and invested in developing with high-NA lith.

TSMC hasn't done that yet.

I don't know how it will turn out, but that strategy sounds ominously like Intel's pre-Gelsinger one. i.e. "We don't need to spend on EUV, we can do it with DUV on a shoestring R&D budget" 10nm/7nm fiasco. Only this time, it's coming from TSMC.
TSMC claims High-NA is too expensive to deploy, just yet. Are they just blowing smoke? We can't know, but if it's TSMC's word against Intel's, I think TSMC has a better track record and I'd tend to believe them.

Keep in mind that, even though everyone is roughly following the same IMEC playbook, there are different ways you can improve a node. It's not as if TSMC N2 is completely non-competitive without backside power, and it's not like their next node will be completely non-competitive without High-NA. They can find other ways to make improvements that might compensate, and might even be cheaper.

Worth noting that TSMC is way behind their original N2 timeline. They slipped by a year.

If you're going to judge TSMC by their roadmap as it was in 2020, then you should also be judging Intel by how its roadmap looked in 2020. It's only fair.
 
  • Like
Reactions: thestryker
TSMC claims High-NA is too expensive to deploy, just yet. Are they just blowing smoke? We can't know, but if it's TSMC's word against Intel's, I think TSMC has a better track record and I'd tend to believe them.
The size of the machines was a real problem for Intel and that was installing the R&D ones. I'm certain it's too expensive for them to install in fabs running older nodes they'd be willing to sideline. I'm assuming their newer fabs were likely built with High-NA size in mind, but since I'm pretty sure they're all running capacity the economics would have to take into account any operational disruptions or delayed startup for new ones.
 
TSMC claims High-NA is too expensive to deploy, just yet. Are they just blowing smoke? We can't know, but if it's TSMC's word against Intel's, I think TSMC has a better track record and I'd tend to believe them.

Well, there's a sucker born every minute.

TSMC, the company that convinced people like you that N7 meant 7nm..

But TSMC's vice president of corporate research, Dr. Philip Wong, was keen to point out that after introducing his company's latest node, despite a history of the node naming scheme actually having some relevance to the silicon features etched into the wafer, the node names are now effectively meaningless. So, while we might like to think that the N7, N5, and N3 names it's using for its 7nm, 5nm, and 3nm nodes relate to the gate length of transistors, they're effectively just brand names.
https://www.pcgamesn.com/amd/tsmc-7nm-5nm-and-3nm-are-just-numbers


Keep in mind that, even though everyone is roughly following the same IMEC playbook, there are different ways you can improve a node. It's not as if TSMC N2 is completely non-competitive without backside power, and it's not like their next node will be completely non-competitive without High-NA. They can find other ways to make improvements that might compensate, and might even be cheaper.

No they aren't 🤣🤣🤣

If you're going to judge TSMC by their roadmap as it was in 2020, then you should also be judging Intel by how its roadmap looked in 2020. It's only fair.

Intel nailed their roadmap:
intel_amd_2021_22_roadmap.jpg