Hacker News Reader

Comment by mschuster91 | Hacker News Reader

> but they need to have a negotiating position with their single point of failure supplier

They don't. Apple has an Architecture License for ARM, they can do whatever the fuck they want.

If you're referring to TSMC - it shouldn't be too much of a problem for Apple to go and contract Samsung instead, at least assuming Samsung can keep up the yield. Intel isn't a competitor to either TSMC or Samsung, their fab process is years behind.

View on HN

Intel’s 18A is closer to availability (functional, ramping to production) than Samsung’s SF2 (still in dev/testing phase); which is roughly analogous to TSMC N2.

TSMC is ahead, as usual, but Intel is closer than Samsung (in this specific case).

“ Such metrics are often closely guarded trade secrets. But according to the Dailian report, Samsung's yields for SF2 are in the 50% to 60% range, just high enough for commercial production. The same report puts TSMC's upcoming N2 node at 80%”

Looks like Samsung is actually closer to production than Intel 18A which is still having issues with yields.

https://www.pcgamer.com/hardware/graphics-cards/samsungs-nex...

Dailian is being overly charitable to Samsung and downplaying Intel. They’re a Korean news outlet with a vested interest in the Chaebol.

That being said, take it however you like. Apple is talking to Intel to make their deal with TSMC more favorable. They could have done the same with Samsung. Either way, TSMC will still be fabbing (at least a good chunk of) their 2nm chips.

> Intel isn't a competitor to either TSMC or Samsung, their fab process is years behind.

It's certainly in Apple (and every company that requires a leading-edge fab)'s interest to try and keep Intel competitive with TSMC and Samsung. 3 companies is already too few for a truly competitive market. And 2 is worse.

I'd argue it's also in everyone's interest to have some redundancy in the chip fabrication supply chain (esp. given the geopolitical situation in Taiwan). It would already be catastrophic if TSMC's production was disrupted for any reason. It would be even more catastrophic if there was no Intel.

Fully agree with you on all points, but I fear this requires serious governmental interventions - simply due to the massive amounts of money involved. The "free market" obviously has failed, with - as you mentioned - massive dangers to national security.

Unfortunately, I also can't see any government willing to put the money on the table to establish a third party from scratch. All that seems to be available is handouts for TSMC to construct a fab in Arizona, and even that was widely criticized.

> I fear this requires serious governmental interventions - simply due to the massive amounts of money involved

Probably. And I suspect it will happen. Chips are crucial, and the governments know it.

Mind you, Apple also has large sums of money available (I suspect more than they know what to do with). So some of that going towards propping up Intel may be no bad thing.

> Unfortunately, I also can't see any government willing to put the money on the table to establish a third party from scratch.

I believe the Chinese government is in the process of doing this with SMIC. It seems likely that they will be competitive before too long.

Everyone’s interest except, perhaps, the Taiwanese state… where TSMC’s know-how lives, and whose help you’d need to transfer it elsewhere.

I keep hearing SMIC has allllllmost caught up, of course…

EUV is an American technology which we decided to license to ASML/TSMC instead of Nikon/Canon. They do have their own knowledge but are not irreplaceable.

Nope, it's Japanese https://en.wikipedia.org/wiki/Extreme_ultraviolet_lithograph...

Nope, it's not. From your own link:

> In 1991, scientists at Bell Labs published a paper demonstrating the possibility of using a wavelength of 13.8 nm for the so-called soft X-ray projection lithography.[4]

> To address the challenge of EUV lithography, researchers at Lawrence Livermore National Laboratory, Lawrence Berkeley National Laboratory, and Sandia National Laboratories were funded in the 1990s to perform basic research into the technical obstacles. The results of this successful effort were disseminated via a public/private partnership Cooperative R&D Agreement (CRADA).[3]

This is a pointless pissing match. It took decades to develop EUV into a high volume tool with R&D effort from different regions that each now contribute to some key part of the tool and supply chain. ASML or Zeiss does not make it a EU technology, the mask and resist does not make it a Japanese technology and Applied Materials does make it a US technology.

Which countries have exercised veto power over sale of EUV technology?

If it was Japanese we'd be talking about Canon and Rapidus instead of ASML and TSMC.

> Intel isn't a competitor to either TSMC or Samsung, their fab process is years behind.

Is it? I've read that Intel's newest process is closer to TSMC N3 than N2, but surely it's not years behind Samsung? I think the biggest problem for Intel right now is acquiring customers and learning how to work with them (but the new CEO should be the right person to do that).

https://en.wikipedia.org/wiki/3_nm_process#3_nm_process_node...

https://en.wikipedia.org/wiki/2_nm_process#2_nm_process_node...

Strictly speaking, Samsung currently has no analogous product to 18A or TSMC 2nm.

Reply on news.ycombinator.com