Oh I very much agree on this point. Indeed, a smaller company might’ve lost such a case by running out of resources to fight rather than on the merits. That can happen in any system of course, but especially if the whole process takes 15 years. To be fair that is I think longest I’ve seen in such a case (Apple’s tax case took 8), but still, not appropriate that it took this long to adjudicate.
The Fall of Intel
- Thread starter Eric
- Start date
While it seems unlikely that Intel will fail completely, it also at one time seemed unlikely that it would fail to the point that lawmakers would hold conversations about what to do if it needed a bailout to survive:
www.tomshardware.com
Intel might be too big to fail — Washington policymakers are already discussing potential solutions if the chipmaker cannot recover
Too big to fail.
An Intel bailout or subsidized takeover would be good, specifically for the fabs. Intel is not far off Samsung/TSMC and those are the only three left in the leading edge, they are an asset for the American (and Western economies generally) economy and national security as well. The fabs have only recently begun a serious move towards openness to third parties and reform, and lacking that was part of what incentivized I think their trip up before before so it’s not exactly crazy to do.
However conditional on the former the design unit should be spun off. They don’t need to be subsidized and shouldn’t be. And in the latter (a takeover) they’d effectively be winded down by QC or whoever and the fat would be cut.
However conditional on the former the design unit should be spun off. They don’t need to be subsidized and shouldn’t be. And in the latter (a takeover) they’d effectively be winded down by QC or whoever and the fat would be cut.
- Joined
- Sep 26, 2021
- Posts
- 6,318
- Main Camera
- Sony
agreed. what’s strategic is the fabs. Intel’s designs could disappear tomorrow, and we’d all happily move on with Arm and AMD, to the extent necessary. But the U.S. shouldn’t let the fab part of the company fail.
For the first time ever, AMD outsells Intel in the datacenter space
But both fall far behind sales of Nvidia's AI GPUs.
AMD outsells Intel in datacenter CPUs for the first time.
I'm not sure if this deserves its own thread or not, but it isn't just Intel struggling with its foundry. Samsung's Foundry is also struggling - both at getting customers and in keeping up with the latest fabrication techniques. It looks like GAA at "3nm" was not a good idea:
www.notebookcheck.net
In terms of long term health for fabrication, just being left with TSMC is a very bad thing for both competition and strategic redundancy.
Samsung Foundry's second-gen 3 nm node yields are reportedly stuck at 20%
Samsung Foundry hasn't had much success in improving yields on its second-gen 3 nm node (3 GAP). They're apparently still at 20%, not enough to be viable for mass production.
www.notebookcheck.net
In terms of long term health for fabrication, just being left with TSMC is a very bad thing for both competition and strategic redundancy.
Not to mention China now has a lot less to deter an invasion of Taiwan…
New class action lawsuit over 13th &14th gen CPU:
Man sick of crashes sues Intel for allegedly hiding CPU defects
Man sick of crashes sues Intel for allegedly hiding CPU defects
One way of saying "elaborate" is "complex". Complex = expensive. After the 10nm fiasco (which as I understand it, in layman's terms was due to being too ambitious to actually make properly) something that is "elaborate" from intel sounds sketchy. I owned a 10nm ice-lake laptop, it was the most disappointing processor I've ever purchased.
AMD gain over 5% of desktop market share in a single quarter and currently has the best gaming desktop CPUs for at least the next year:
www.tomshardware.com
Gelsinger’s previous comments about TSMC not being reliable due to geopolitics lost Intel an important discount on TSMC products.
www.tomshardware.com
His latest comments talked about how awesome TSMC is and what important partner it is and how Intel also supplies tools to TSMC.
AMD's desktop PC market share skyrockets amid Intel's Raptor Lake CPU crashing scandal — AMD makes biggest leap in recent history
Though Intel says it was an inventory adjustment at one of its customers.
Gelsinger’s previous comments about TSMC not being reliable due to geopolitics lost Intel an important discount on TSMC products.
Intel CEO reminds investors it provides TSMC with chipmaking equipment — Gelsinger emphasizes that TSMC is a strategic partner
Well, IMS Nano does.
His latest comments talked about how awesome TSMC is and what important partner it is and how Intel also supplies tools to TSMC.
- Joined
- Sep 26, 2021
- Posts
- 6,318
- Main Camera
- Sony
AMD gain over 5% of desktop market share in a single quarter and currently has the best gaming desktop CPUs for at least the next year:
AMD's desktop PC market share skyrockets amid Intel's Raptor Lake CPU crashing scandal — AMD makes biggest leap in recent history
Though Intel says it was an inventory adjustment at one of its customers.
Gelsinger’s previous comments about TSMC not being reliable due to geopolitics lost Intel an important discount on TSMC products.
Intel CEO reminds investors it provides TSMC with chipmaking equipment — Gelsinger emphasizes that TSMC is a strategic partner
Well, IMS Nano does.
His latest comments talked about how awesome TSMC is and what important partner it is and how Intel also supplies tools to TSMC.
Fire Gelsinger. Seriously. He is not a CEO.
As far as I can tell, every move Intel has made has been to make his long term position more secure. Intel even got rid of one its own rules about age to keep him - about 3-4 years ago they removed the hard limit of mandatory executive retirement at 65 as Gelsinger was 60.
Intel resumes free coffee and tea for its employees — $100-million-per-year program partly reinstated to boost employee morale
Fresh fruits remain out of reach in the breakroom, though.
Well that’s something
Can’t have everything …
AMD gain over 5% of desktop market share in a single quarter and currently has the best gaming desktop CPUs for at least the next year:
AMD's desktop PC market share skyrockets amid Intel's Raptor Lake CPU crashing scandal — AMD makes biggest leap in recent history
Though Intel says it was an inventory adjustment at one of its customers.
Intel claiming that their poor gaming performance is due to optimization and tuning issues and that they totally saw better performance on their internal testing than 3rd party reviewers did out in the wild.
Intel announces Arrow Lake fix coming within a month — Robert Hallock confirms poor gaming performance is due to optimization issues
A fix is on the way.
They’ll have a fix for these issues later in the year. I somehow doubt that they’ll catch up to AMD, but let’s see if they improve (beyond burning power consumption for performance).
Of course on thing that almost always amuses me wrt to CPU gaming reviews is that in many cases, the way gamers actually play their games, the performance is GPU limited and the CPU performance is almost after thought. This isn’t true in all cases of course, but … it is in many if not most of them. So the reviews (and CPU company press release) will wax lyrically about an amazing X% performance increases at 720p lowest quality … while at Ultra 1440p or 4K the difference is zilch.
Samsung claiming that they can’t drive workers hard enough to keep up:I'm not sure if this deserves its own thread or not, but it isn't just Intel struggling with its foundry. Samsung's Foundry is also struggling - both at getting customers and in keeping up with the latest fabrication techniques. It looks like GAA at "3nm" was not a good idea:
Samsung Foundry's second-gen 3 nm node yields are reportedly stuck at 20%
Samsung Foundry hasn't had much success in improving yields on its second-gen 3 nm node (3 GAP). They're apparently still at 20%, not enough to be viable for mass production.
In terms of long term health for fabrication, just being left with TSMC is a very bad thing for both competition and strategic redundancy.
Samsung: Exynos problems are due to "too short" 52-hour working week in Korea
Samsung Exynos chips have not been able to keep up with the fastest chips from Qualcomm, MediaTek and Apple in the high-end segment for years. Samsung reportedly views South Korea's workers' rights as one of the reasons for this, as employees are officially only allowed to work a maximum of 52...
www.notebookcheck.net
Workers saying more hours per week don’t necessarily lead to better results.
Hmmm … Samsung’s Taiwanese fab competitor has even tighter labor rules (as does one of their chip competitors) and their US chip competitors have more people. Meanwhile their executives’ solution is seemingly that the beatings will continue until morale improves. Hiring better people or at least more of them to spread the load is obviously off the table. Or maybe the problem is with the decisions from executives … nah … it’s that the workers aren’t working hard enough.
Hey intel:
the software that exists in the wild is what we run, not some fantasy optimised code in your lab
Makes it even funnier that AMD are kicking the shit out of them in performance at the moment, when most software is still no doubt compiled with Intel's "de-optimise AMD" compilers.
iPhone 18 series: Rumor claims 2026-bound A20 Pro chipset could be built by Intel
Apple's iPhone 17 series is expected to be powered by the A19 chipsets when it arrives next year. Said chipsets are slated to be built by TSMC but a new rumor now claims that the A20 SoCs, bound to launch with the iPhone 18 series in 2026, may ditch TSMC in favor of Intel.
www.notebookcheck.net
Dubious to say the least. We’d almost certainly see this in TSMC’s (upcoming?) financials given that Apple’s huge payments to them in advance to reserve node space is well known. Also apparently TSMC’s own N2 node is going to be seemingly ready on time.
TSMC's says 1.6nm node to be production ready in late 2026 — roadmap remains on track
No, A16 is not delayed, says TSMC.
So we have a long time partner that appears to be delivering its services on schedule vs one that doesn’t have a track record with any customer on a node they themselves decided not to use supposedly because they didn’t want to spend the money to build out volume to focus on the 16A node. And yet this node with no volume production capabilities is the node they are supposed to be building iPhone chips with in a year’s time for a launch in two years? Hmmm … I’d have less difficulty believing it if Apple was testing out Intel fabs with a smaller order of a more limited run chip, but the iPhone?
The leaker themselves seems new - they did correctly predict some new colors on the iPhone16 (though MacRumors said they said 7 colors rather than 5 for iPhone 16 - unclear if so or they just said two new colors) and the bronze color for the Pro. Most of their other “leaks” are pretty obvious or far enough in the future as to not be known. Basically no track record, especially for something like this which would represent a massive industry shift.
Btw one interesting part about the Tom’s article is the implication that TSMC’s A16 might not be suitable for phones or even tablets/PCs:
Tom’s has a follow-up on this with TSMC slides:
TSMC 1.6nm update: Tangible improvements, but new challenges emerge
Best suited for AI and HPC processors.
So it’s possible that Apple might go with N2P for its chips?
I know little to nothing about this, but given their track record so far I can easily believe Tom’s is misapprehending what a TSMC engineer actually said. That said, TSMC does seem to suggest that A16 might only be viable for larger chips “with dense power delivery networks”. Which I’ll be honest I’m not sure what that means for its use in Apple products, but it doesn’t sound conducive for building A/M chips.
Similar threads
