I thought it was Stussersaurus? You know, like a dinosaur, good enough for a simpler time. What’s that word i’m looking for? Obso-something. Elite? It’ll come to me.
Stusserdramus, seer of seers, predicted on HRose’s blog that you’d be a dick today.
Oh, yeah, Intel just ousted their CEO.
Fortunately all swans belong to the Queen so he should be ok.
It might be a good time to buy Intel. I mean a week ago.
That’s the best news I’ve heard about Intel in more than a decade. I knew Pat pretty well when I was at Intel. He was the boy wonder (I think he graduate college at age 20 or so) , in a company with a lot of smart folks it was very rare room, he wasn’t the smartest person in it.
Edit: The statement by Pat at the bottom of this story is good start. Turned out he stated at Intel at age 18.
https://www.anandtech.com/show/16419/intel-appoints-pat-gelsinger-as-new-ceo-from-feb-15th
So, why is Intel having such trouble updating their chip production?
It’s an incredibly difficult technical challenge to continue to drop fabrication sizes these days. Five nanometers is incomprehensibly tiny. A human hair is 100,000 nanometers thick.
It’s really tough to comprehend how small these processes are. I swear I remember hearing years ago that there was a theoretical limit to how small you can go because electrons are too big and they’d start jumping their lanes, or something like that.
This doesn’t seem to have aged well.
Intel Abandoning Silicon With 7nm and Beyond - Silicon Alternatives Coming By 2020 (wccftech.com)
The fundamental flaw with Silicon transistors is that at the 7nm point the transistors sit so close to each other that an effect called quantum tunneling occurs. This effect unfortunately means that the transistor cannot reliably be turned off and for the most part will stay on.
However there’s one very promising short-term Silicon alternative that will most likely supersede Silicon for a few years. It’s a III-V semiconductor based on two compounds and four different elements. Indium gallium arsenide ( InGaAs ) and indium phosphide (InP). Imec, a research center tasked with finding the next thing after silicon, has already managed to fabricate FinFET transistors using InGaAs and InP on a 300mm 22nm Silicon wafer a year and a half ago.
If i were President i’d be pushing for a 3D silicon revolution, like the Human Genome Project sort of science advancing program, since it will take a revolution not just manufacturing but an entirely different kind of programming at the most fundamental level, much more complicated, and probably based on some kind of really insane to understand nodal mathematics.
A silicon atom is 0.2nm thick. So 3nm, which will be coming out next year, is very much on the atomic scale. The connections are only 15 times the size of the atoms of which they’re comprised.
Just because the process is “3nm” it doesn’t necessarily mean the connections are 3nm wide, I thought? (So the intel processes are actually “smaller” than the TSMC processes with the same name).
It’s crazy that basically all modern CPUs could end up being made in a single set of fabs in some of the most disputed territory on earth…
intel are still like 3rd after TSMC and Samsung. And we clearly have under-cacacity in fabs atm.
I don’t think the US would allow that. Intel are like, um, Boeing…
Is USgov going to buy $5bn+ of obsolescent chips off intel a year from 2022/2023? Because unless intel can get their 7nm process (roughly equivalent to the TSMC 5nm process that is working today) off the ground the rest of their fabs are going to be looking obsolete by that point.
I guess it’s very possible TSMC stall out at 5nm and intels 10nm process stays, if not competetive, then at least adequate.
The physical limitation of Moore’s Law have always been 5-10 years away since I joined AMD in 1981. Yet somehow the clever engineers figure out a way to get higher transistor counts, faster speed, and lower power consumption, with each generation. The physics of how this is done have been beyond my comprehension since the mid 90s.
@Alistair is right there are national security implications of this, but I don’t think there is any prayer of getting support at national to fix the problem. In the mid 1980, Japan semiconductor companies were dominant the industry, and Congress approved the funding of Semitech, to help US firms remain competitive in semiconductor manufacturing. It was only marginally useful. But Intel took some of the tech (and lot the invented themselves) and rose from the #7 largest chip company to very dominant #1.
Today the Japanese are very much also rans in the chip business. So fortunates can and do change.
Pat’s only one guy, but there is some hope, especially if you look at AMD rise under the leadership of a real engineer like Lisa Su.
The EU seems to be planning to massively subsidise new fabs and R&D. I can see the US doing the same.
Foreign supply chain dependencies are at the top of responsible leaders’ minds these days after PPE shortages throughout 2020 cost lives.
Not sure what the right answer would be for chip fabrication and electronics, as the blocker isn’t only funding but also engineers. How do you reduce reliance on China and Taiwan?
The obvious answer is strong tax incentives combined with already-existing tariffs to incentivize companies to fabricate and assemble locally, but again, we just don’t have enough engineers, and there’s no Shenzen equivalent, a massive metropolis full of young people building electronics. It would probably take a generation to catch up to China in mass production of consumer goods, and it would be incredibly expensive paying US wages to do it. The incentives and tariffs required for that to make financial sense would be burdensome.
I worked in the computer chip industry back in 2004-2007, with Intel, AMD, tsmc, Charter, TI, IBM, Samsung etc. as customers to our process chemistries. I remember talking with guys at Intel and in conferences about the huge challenges of getting to 45 nm, and how 22 nm may be theoretically impossible due to the inability to have sufficient dielectric layers and the calculated heat the processors would generate. I couldn’t imagine 10 or 5 nm.
Of course, I’d have never imagined TSMC producing Intel chips. I’ve been to their fabs in Taiwan many times. Intel’s rep was they could produce chips that no one else could make, or at least no one else could reproduce their process. In fact, you’d see some companies in, e.g., the CMP business who would get lured into the Intel trap of “If you sell us your product at your cost, you’ll be able to sell it to everyone else because you can tell them Intel uses it.” The problem was, no one else’s manufacturing process was like Intel’s so some things that worked for them didn’t work for anyone else.
I will say I am really happy to be out of that business. I’ve led R&D in a lot of markets for a variety of companies, but that one was the most… interesting.
Hopefully they continue to build facilities elsewhere. If China ends up being the only other country with any real domestic chip production then that seems like a strategic nightmare for the US and allies.
Maybe in 10-20 years we’ll know the details of what happened to Intel and why they fell behind.
I do think we’re getting close to the limit of 2D semiconductors - i mean, like stusser said, we’re reaching atomic sizes. There’s… not a lot of “manufacturing capacity” for sub-atomic semiconductors, so to speak.
It’s not China, it’s Taiwan. Very different place-- but it is contested.