News Intel CEO will reportedly present plans to cut assets at an emergency board meeting — chipmaker may put $32B Magdeburg plant on hold and sell off A...

[Admin]

Intel CEO Pat Gelsinger is expected to meet with the board of executives this month to share potential plans on what to cut from Intel to turn around the company. Intel's Foundry wing will not be on the chopping block, but Intel's in-progress Magdeburg fab will likely be canceled.

Intel CEO will reportedly present plans to cut assets at an emergency board meeting — chipmaker may put $32B Magdeburg plant on hold and sell off A... : Read more

 

[Neilbob]

But hey, at least AMD is in the rear-view mirror. Whee!

Sorry, I couldn't help it. But it is an appropriate reminder to never, ever take the statements of a CEO too seriously. ANY CEO.

 

[ekio]

I hope the corporate will have the decency of decreasing their salary by a factor of at least 10, their work has been such a failure. They should even quit on their own an apologize to the staff.

 

[-Fran-]

Bean counters needing that short term money to bail out the shareholders, so they can cash out and shorten Intel into nothingness. Well, that could be a potential extreme, but that's how it would roll if Intel didn't have anything left, which I'd say they have plenty left to survive.

Sad to see Intel being stripped appart to cover the incompetence of the current c-suite managers at the company, but that is what happens when you have no real leadership for a while. After Otelini left, it was pretty much every person on their own, it looks like. Zero corporate responsability and accountability. I mean, just look at BK's exit. That was a real crapshoot.

Regards.

 

[hotaru251]

to be fair..any mega corp on top 500 list should be required to slash the top heads pay before selling off/fireing ppl in mass.

You should never reward the top brass for their failure in leading a company properly.

 

[Thunder64]

Good thing they did all those stock buybacks rather than invest that money into R&D or their workers, or fabs? Anything really.

 

[bit_user]

Good thing they did all those stock buybacks rather than invest that money into R&D or their workers, or fabs? Anything really.

Yeah, they simply inflated a stock price that they're now having to take desperate measures to prop up. If they'd done as you said, then the stock would've languished more in the short term, but the long-term health of the company would be better.

IMO, the vesting schedules of executives' stock options needs to be longer, so they take a more long-term focus on the company's health. Boards will never do that, because boards represent investors' interests and investors are too short-term focused, these days. It would probably take the SEC to issue such a ruling.

 

[Dustyboy1492]

Seems necessary, the foundry side has more employees than TSMC and the design side has more employees than AMD or Nvidia, something needs to give. Cut the side projects and get lean, focus on the core business and bringing up the new fabs.

 

[DavidLejdar]

If I were on Intel's board of directors (such as appointed by an emergency shareholders' meeting), I would remind everyone there, that by end of June, year-on-year revenue was up 1.99%, compared to the previous year. And $55B may not seem much, compared to the $79B at the end of 2021. What is interesting though, is that $55B is as much yearly revenue as at the end of 2015. Yet, while at end of 2015 the gross profit was $34B, at the end of last June it was less than $23B.

And before hitting an eject or self-destruct button, that's something that would seem worth a closer look at. Like, did the production costs increase heavily, and that perhaps due to inflation? Did U.S. import duties, and/or export restrictions, have an impact on business? I.e. when Intel China had import/export with Intel elsewhere, did the like company supply lines get affected? And stuff like that.

After all, if Intel wants to stay global, that requires some attention. You know, instead of perhaps just relying on that AI is going to carry the day, despite not really (yet) offering that a strong argument about why one would want to upgrade e.g. to a new laptop (or buy the more expensive one), just because of some MS Copilot feature, when the old laptop just works fine for the few office applications, and in a climate of layoffs at that, while these layoffs may perhaps not be that necessary, when there would be more focus on the business-side of the business.

I.e., people like practical, right? And how practical would it be, if there was a long box (with CPU, iGPU, NVMe), that could easily be plugged in e.g. behind the screen from above? I mean, that may perhaps get a perplexed reaction from a focus group (in market research). But, not every desktop user is necessarily a gamer, or is a gamer, but not at work. And thanks to NVMe, less space needed. So, what could technically be offered, is that one doesn't need more space than e.g. for an 27- or 32-inch screen, to be able to run a desktop setup (with iGPU) - and with less cables at that... just one power-plug, and USB ports at the foot of the screen.

 

[bit_user]

Seems necessary, the foundry side has more employees than TSMC

Has TSMC surpassed Intel's wafer volume, yet?

Also, Intel has done their own in house EDA tools, whereas TSMC works with partners like Synopsis. It's possible TSMC also outsources other things that Intel does in house, so the numbers might not be exactly comparable.

the design side has more employees than AMD or Nvidia, something needs to give.

Is that counting Altera and MobilEye? Also, exactly when were these numbers taken? For AMD's part, do they include Xilinx?

Cut the side projects and get lean, focus on the core business and bringing up the new fabs.

I've only heard of Gelsinger cutting projects and selling things off, since he came on board. They haven't gotten into any new businesses or started any major new projects, since then. That's why they're at the point of having to dump assets, now. There's not much more they can cut which doesn't affect their core strategy.

 

[vanadiel007]

Well, there's always the option to form a joint venture with AMD and compete against Nvidia.
I bet nobody would have seen that one coming.

 

[bit_user]

If I were on Intel's board of directors (such as appointed by an emergency shareholders' meeting), I would remind everyone there, that by end of June, year-on-year revenue was up 1.99%, compared to the previous year.

The industry was still in a slump, 1 year ago. You'd naturally expect revenue to be up. The thing to do is compare with the market, overall. If they're not growing as fast as the competition, that's not sustainable.

before hitting an eject or self-destruct button, that's something that would seem worth a closer look at. Like, did the production costs increase heavily, and that perhaps due to inflation? Did U.S. import duties, and/or export restrictions, have an impact on business? I.e. when Intel China had import/export with Intel elsewhere, did the like company supply lines get affected? And stuff like that.

Good questions, but they're not n00bs. I'm sure they understand their business well enough to know what changed and how the current trends are likely to play out.

how practical would it be, if there was a long box (with CPU, iGPU, NVMe), that could easily be plugged in e.g. behind the screen from above?

You mean like the VESA adapters that let you attach a mini-PC to the back of a LCD monitor? Those have been a thing for like 10-15 years and the mini-PC market was developed by Intel via their NUC product line (which Gelsinger sold off to ASUS, BTW).

 

[bit_user]

Well, there's always the option to form a joint venture with AMD and compete against Nvidia.
I bet nobody would have seen that one coming.

That would flout anti-monopoly and anti-collusion rules. 0% chance of it happening.

Plus, with AMD on the rise, what would be in it for them? Why would they want to weigh themselves down by having to carry Intel?

I guess the closest they came to this was the Kaby Lake-G product, where Intel integrated a Vega (actually Polaris?) die on-package, in a laptop processor (with HBM for the GPU). That was in like 2018, before Intel had their own dGPUs. But, AMD was still tiny and struggling, back then. Not the competitive threat they are, today.

 

[JRStern]

Would Gelsinger like to return about 90% of his compensation since rejoining?

While I'm sure these moves are well justified, I can't help suspecting there is some big piece of bad news that's triggering it now.
Some expected progress will be pushed off, some big orders cancelled, some big competition about to announce. Something.
Yet the stock was *up* on Friday on these rumors.

 

[NinoPino]

Well, there's always the option to form a joint venture with AMD and compete against Nvidia.
I bet nobody would have seen that one coming.

I think it could be possible only on the fab side of business, and only if Intel decide for splitting production, and if AMD have enough money to invest.

 

[t3t4]

Well my i9 13900k is still working just fine! But I was smart enough to know that 4000 watts will smoke the thing, so I never let mine run unlimited. Now I wait patiently to load up on Intel stocks since everyone sold off .

 

[jp7189]

Arrow/Lunar (fabbed at tsmc) will be telling. For the first time Intel will have a direct comparison of their fab vs tsmc. If their design on tsmc does well, it might save the company, but it might make it harder to bring silicon back to Intel foundary which would then have to be spun off.

 

[jp7189]

Has TSMC surpassed Intel's wafer volume, yet?

Intel's GPUs and now CPUs are fabbed at TSMC. I'm not sure if Intel fabs are still 100% utilized, and if not, that extra capacity would be a liability.

 

[vanadiel007]

That would flout anti-monopoly and anti-collusion rules. 0% chance of it happening.

Plus, with AMD on the rise, what would be in it for them? Why would they want to weigh themselves down by having to carry Intel?

I guess the closest they came to this was the Kaby Lake-G product, where Intel integrated a Vega (actually Polaris?) die on-package, in a laptop processor (with HBM for the GPU). That was in like 2018, before Intel had their own dGPUs. But, AMD was still tiny and struggling, back then. Not the competitive threat they are, today.

Well, I would not be so fast with stating it will fail due to anti-monopoly and anti-collusion rules.

I would make the argument a joint venture with AMD would be better for consumers than Intel going bankrupt and their IP being sold off to the highest bidders, giving AMD a de facto monopoly position in the consumer CPU market.

 

[bit_user]

Arrow/Lunar (fabbed at tsmc) will be telling. For the first time Intel will have a direct comparison of their fab vs tsmc.

Yes and no. The N3B node I believe they're using is already like a year old, while Arrow Lake would be the first thing on Intel's 20A node. So, it won't be exactly an apples-to-apples comparison. Still, it should be interesting.

IIRC, 20A will use gate-all-around transistors and backside power delivery, neither of which are present in TSMC's N3 nodes.

 

[bit_user]

I would make the argument a joint venture with AMD would be better for consumers than Intel going bankrupt and their IP being sold off to the highest bidders,

Who says Intel is on the brink of bankruptcy?

If the joint venture involves something like Intel spinning off their fabs, then I guess AMD could take a stake in that newly independent foundry. I'm not sure they'd want to, after finally washing their hands of Global Foundries.

FWIW, I'd point out that Intel's join venture with Micron, involving 3D XPoint memory (i.e. Optane) didn't work out well. When the partnership unwound, Intel asserted ownership over all the IP, leaving Micron with none of it and a production line they couldn't use. Intel was also left without a fab for Optane, but the future of the technology looked unclear for other reasons.

 

[jp7189]

Yes and no. The N3B node I believe they're using is already like a year old, while Arrow Lake would be the first thing on Intel's 20A node. So, it won't be exactly an apples-to-apples comparison. Still, it should be interesting.

IIRC, 20A will use gate-all-around transistors and backside power delivery, neither of which are present in TSMC's N3 nodes.

Will there be a variant of Arrow on an Intel fabbed process node? I'm hearing Panther will be the first to return to Intel Foundary and only if they have a competitive process at that time.

 

[Phyzzi]

Would Gelsinger like to return about 90% of his compensation since rejoining?

While I'm sure these moves are well justified, I can't help suspecting there is some big piece of bad news that's triggering it now.
Some expected progress will be pushed off, some big orders cancelled, some big competition about to announce. Something.
Yet the stock was *up* on Friday on these rumors.

In this case, it's not really a secret. AMD has CPU's already that are just as fast for most things, use half the power, and cost a similar amount or less, as well as often being able to use the same MoBo for a few generations. The Radeon 7900X easily competes with the 14900KF and that's before they add VRAM. Given that the 7850X3D also runs about as fast as anything Intel can put on the table, the inevitable vram upgrade to this generation of Radeon CPU's is going to make this one a real beast for gaming, and pretty solid for plenty of other things, while still almost certainly running cooler and lower powered than anything Intel can push. AMD also has a pretty good lock on the midrange and lower range CPU market, again with competitive pricing and efficient CPUs. Further, Intel trying to jump on the fast moving train of discreet GPU's has been a laughable failure. With driver issues, poor performance, and VERY serious competition, they are barely worth mentioning. That, and losing Apple and finding ARM to be competitive even beyond mobile devices means Intel is no longer the clear best choice in any of its former solid markets.

That doesn't mean they are totally out, but they are going to have to refocus, reprice, and find ways to be a better value than AMD. Jumping on AI is not stupid, but it won't make the difference alone. In PC's they certainly can't continue to hold the market on single core speed alone, in servers they need to have something that does a better job of scaling at high value, and in mobile they need to make serious gains in efficiency. They are being smart in opening up their fabs to other designers: even if TSCM has managed to stay ahead by breaking through 10nm well ahead of Intel, there are plenty of chips that don't need to be made on the very limited number of fabs that can do work that small. Intel should also probably get into their own SoC work a bit more assertively. Beyond just competing with ARM more, having x86 be able to run efficiently in crossover mobile environments is clearly one of the emerging markets. If they can manage to make some inroads in new directions, they can still be a force in the market, even if they probably can't ever retrieve the total dominance they have had before in some markets.

That said, the rumor mill is that Intel is cancelling projects. That may be evidence of them acknowledging that certain pursuits aren't likely to pan out and an acknowledgement that they need to change direction, but it also is just undeniably a loss of potential in some markets.