Yes. I can give you a non-technical answer, since HN is ostensibly business as well.

Intel fired the one CEO that spoke both engineer and business, and Gelsinger could have been their Lisa Su. They fired the only talented CEO they've had for years.

This will be fatal.

Gelsinger was the scapegoat for 20+ years of inability to compete with foreign companies, no matter how much money was poured into them. They used American exceptionalism as a cover to defraud shareholders and any government that invested in them. They used the relationship of AIPAC and Congress to build a fab and R&D lab in Israel (inserting yourself into global politics to make a buck is always spicy) at low cost to them.

Taiwan became the capitol of electrical engineering in the world, and is a shining example of how to survive and thrive in a post-war era, and it absolutely shows. They caught up to Intel and zoomed right past.

Gelsinger's crime was try to do what AMD did: they didn't have a fab that could make their chip BUT they had a fab that made chips that people wanted AND the foundry could take that work and survive if they legally split. GloFo is now the third largest semi foundry in the world today, and when it was part of AMD, it very much wasn't; I can't quite remember, but 5th or 6th? Something like that. GloFo is #3, TSMC is #1, Samsung is #2, and Intel could very well be that #4, and push out UMC (#4) and SMIC (#5) in the secondary chip foundry market.

Gelsinger could have split Intel into Intel and IFoundry or something, and Intel could have profited on IFoundry taking off and taking external work. Right now, IFoundry can't compete on top nodes, but _could_ steal work from all other fabs for secondary larger nodes. Having a working 12 nm competitor as well as a working 7nm competitor is big business, which Intel currently has _ZERO_ of (since they don't take external contracts). Gelsinger was big on this potential revenue stream.

Gelsinger's other crime was being part of the negotiation between TSMC and the Biden administration for the CHIPs act money: part of what built the TSMC fab right next door to Intel's in Arizona was Biden and Intel money. Intel was investing in it's future by playing the American exceptionalism card again, but now in everybody's favor. We _all_ benefit from this. Gelsinger wanted to have _somebody_ fab the chips, and if its good enough for AMD, Apple, and Nvidia, its good enough for Intel.

There is zero indication that GAA 20A is ready, and Intel has a history of having leadership that says such-and-such is ready for it to either come out several gens later, or just vanish off the roadmap. Gelsinger's other OTHER crime is admitting to this and changing the direction of the Titanic before it hits the iceberg, for the CEO that replaced him just to steer right back into the iceberg.

I have _zero_ faith in Intel's leadership if they can't bring Gelsinger back. Tan, Gelsinger's replacement, is a former board member. I have no reason to think he is not just going to further poison the company. Tan has not spoken about any plan that indicates he understands Intel is not competitive, Intel cannot competitively make 100% of the tiles, that Intel's Foveros tech stack is extremely valuable because the only truly comparative alternative is TSMC's CoWoS tech family and superior to it and people are willing to throw money at that problem but they can't license it as long as IFoundry is part of Intel.

Intel is cooked imnsho.

▲

FuriouslyAdrift 5 days ago | parent | next [-]

Intel has been in Israel since 1974. Intel Fab 8 was built in 1980 in Jerusalem... There's over 30,000 chip engineers and nearly 200 semiconductor companies there, now.

▲

DiabloD3 5 days ago | parent [-]

AIPAC was founded in 1954.

▲

FuriouslyAdrift 5 days ago | parent [-]

Intel came to Israel mostly because Dov Frohman (one of Intels first employees who had worked with all the founders at Fairchild and also the inventor of EPROMs) pushed to establish an Intel dev center there when he moved back home.

At the time, EPROM tech was Intels most profitble product until the 8088 and 8087, which were designed in Israel at the dev center (along with many of their chip designs).

	▲	DiabloD3 4 days ago \| parent [-]
		Yep, Dov Frohman's contribution to tech is well known and very appreciated. Edit: Look, to whoever is out there on a downvote spree, I don't care if I get downvoted, man, but wild you'd just downvote people talking about a guy whose won multiple IEEE awards, has patents to his name, and has left his mark on EE, and isn't even the focus of the discussion at hand.

▲

isthatafact 4 days ago | parent | prev | next [-]

I am no expert in Intel, but in my view, Gelsinger lost the faith of many by being unrealistically optimistic. Of course a CEO needs to be optimistic, but he promised (in 2021) zettaflop systems by 2027 (the worst example I remember). Did anyone believe that could happen?

His over-optimism gave the whole "5 nodes in 4 years" supposed path to leadership a weird flavor, like it must be somehow a bit of a con even if it gets technically achieved.

	▲	DiabloD3 4 days ago \| parent [-]
		I mean, if I thought I had a plan to be the guy who saves Intel from it's own mistakes, I'd be optimistic too. Also, I looked into the claim when he had said it, apparently he was being intentionally misleading about it, and the press tried to ask what he meant: he was speaking tensor performance on future enterprise Arc card products at datacenter scale, ie, AI bait. In early 2021, Nvidia's compute flagship was the A100, 19.5 FP32 MAD TFLOPs, but the misleading number they quote in marketing is the tensor performance of 312 TFLOPs of FP16 accumulates. That would be about 3.2 million of these at tensor perf. Skipping H series, in late Nov of last year, their new flagship is the B200. 124 FP32 MAD TFLOPs, 2250 tensor FP16 accumulate TFLOPs. That is now 445k cards to reach zettascale if using tensor cores. You won't be fitting ~1400 GPU-laden machines in a single datacenter, but the number is becoming more manageable. They improved, in 3.5 years, 7.2x. Lets say Nvidia does this again. 3.5 years, again, would put you in early 2028, and they manage another 7.2x win: that could be 62k cards across 7.7k. That absolutely is doable in a single datacenter. The problem is, and this is where the prediction actually falls apart, not that its impossible: We don't know what future Arc cards look like, nor enterprise ones. Battlemage is an improvement over Alchemist, so the tech *is moving forwards at, but either Celestial or Druid was supposed to introduce the enterprise compute card variants, but that seems to be dead, and no indication either of those lines will even see the light of day now. The new CEO seems to be hard set on making Xe for iGPU only. I can't find any hard numbers on Intel's tensor units, but apparently they're actually competitive. I can find the normal FP32 MAD numbers, and it ends up that Intel is 13.5w per TFLOP and Nvidia is 8 and both companies have equal efficiency in transistor usage. Assuming Intel made a B200 competitor, and assuming the higher power usage is due to voltage (Intel B series voltage is similar to Series 40's voltages, which is a lot higher than equivalent enterprise/pro series cards), Intel could be making a card that's somewhere in the ballpark as 2/3 as good for the same power usage. So, in the end, yes, I don't agree with his claims of future Zettascale at Intel by 2027. I don't think he was wrong for the industry as a whole, however. If he would have said, say, 2030, I don't think we would be discussing this, that certainly would have been doable if he was at the helm and they kept doubling down on Arc every gen and everything went according to plan.

▲

1718627440 4 days ago | parent | prev [-]

Why needs Intel to split in order to make contracts with other companies? Can't they just do it when they are still a single company?