How does Govt picking winners and losers going to help?
Intel is no Too big to fail Bank. Why save Intel of all chip manufacturers? Wouldnt it be like 25 years too late, with Intel and its heydays !?
Would Govt now ensure parity by investing in "marquee" entities across different industrial domains?
By ensuring that the US retains at least the ability to manufacture second tier CPUs vs complete reliance on Asia? This doesn't seem unreasonable.
AMD no longer has a fab. TSMC dominates the global market and basically has no competition.
In the event that Taiwan is invaded, the US would suddenly have a huge problem getting access to any kind of high end chips, be they CPUs or GPUs. This would be a major problem economically and militarily for the US.
Some caveats: Due to the chip act, TSMC does now have fabs Arizona, though I'm not sure what their capacity is. TI, and some others building lower end components also have fabs I believe. For x86, high end ARM, and GPU's, virtually all of that is manufactured by TSMC right now, mostly in Taiwan.
It's the only chip manufacturer "left" in the US. The argument is national security: the US expects China to invade Taiwan and this will kill TSMC in the process.
Whether this will happen or not can be debated, but this is what the government expects.
But if all of our advanced weaponry used chips from Taiwan or Korea, for example, then the strategic implications for war in East Asia would be radically different. People are right to say that China could engage in war over Taiwan for chips, but for the wrong reasons. It's not that they want access to the fabs (they'd love it, but they're not stupid and they know the fabs and know-how would be destroyed in the war), but it would deny the US defense industry access to those fabs.
If US missiles or drones use chips from TSMC, and TSMC is in occupied territory or a war zone... the US can't make more missiles or drones. And no matter how powerful your starting position is, you can't wage war without the ability to replenish your stockpiles. It's the bitter lesson Germany learned in both world wars.
China wants hegemony in Asia, and to remove the influence of the US, Japan, and their allies within what they perceive as their exclusive sphere of influence. How to achieve that? Invade Taiwan, which eliminates western access to TSMC one way or another, effectively blockading western defense industry from the core things they need to resupply their militaries in a war. Like WW1 all over again, a "preemptive war" becomes the game-theoretic optimal outcome, and the world suffers.
How to counter that? The US and its allies need to make sure they have access to chip fabrication facilities that can produce near-state-of-the-art chips, even at inflated prices that are not commercially viable in peacetime, as well as the necessary strategic minerals like germanium and lithium. Only then does calculus swing the other way in favor of peace. Hence Biden's effort to get TSMC to build SOTA fabs in Arizona, and when that failed/stumbled, this investment in Intel.
Global Foundries, Micron, and Texas Instruments all come to mind
As a software engineer, this isn't an entirely new concept.
Would it though? The TSMC foundries are pretty much in every continent. Are they just going to stop operating if this happens? Because that seems akin to killing a golden goose.
Also what is up with Global Foundries? I don’t hear a peep about them.
I think if this was a domestic thing it would be all kinds of dumb and wrong. But as a US National Security thing, it makes sense if you’re of the mind that significant intervention is fine when it’s in your country’s best interest.
The next phase is watching the U.S. government keep Intel on a palliative drip of softball contracts and tax dollars. I guess there’s a fair argument that this form of bail out could help Intel thrive again… or at least secure a domestic supply of chips for natsec reasons?
With Intel maintained, if China invades Taiwan and takes TSMC the US will still be able to make usable processors. They won't be the latest and greatest like TSMC, but they will be good enough. Maybe not the most powerful or efficient, but still rather close.
My only worry is this will mean management will start resting on their laurels and things will just continue to deteriorate. Or maybe the government can convince them to get rid of the bad management and start thinking more long term and less about immediate profits.
It's a terrible idea
wtf? what do you mean, they're like less than 1 year behind TSMC when it comes to leading node
Edit: I think it's a misconception that China cares much about fabs in Taiwan. It wants unification.
180,000 wafers a year. Globally they do 17 million. They announced first profit yesterday.
Um.
All that stuff is still semiconductors, just with different patterns printed on them.
The thing about drones is that they actually don't require much computational power compared to modern consumer computing. It's just math - control systems, calculus, trig, waypoints, etc. All of these were solved problems in the days of the Apollo Guidance computer, and will run comfortably on chips from 2 decades ago. The STM32F722 microcontroller that is one of the most common hobbyist drone chips is built on the 90nm process node, runs at 216MHz, has 512K of SRAM, and costs about $5/chip. FWIW, it's made in France and Italy rather than China, and STMicroelectronics owns its own fabs rather than outsourcing to TSMC or Chinese companies.
If you want to do things like computer vision on the drone, the computational requirements are quite a bit higher, but you can still run something like YOLO at orders of magnitude less computational power than what you've got in a Pixel 9 or iPhone 16.
...which makes me wonder if a better strategy for the military would be to fund a wide variety of domestic chip manufacturers operating at decades-old process nodes (eg. the 65nm process node from 2005 seems to be at about the sweet spot), rather than try to prop up the one American company that can compete on cutting edge 7nm process nodes. Particularly since the experience of WW2 was that simple, robust designs that could be easily licensed to other suppliers and mass produced (eg. the Hawker Hurricane, Grumman F6F Hellcat, Grumman/GM TBF Avenger, Liberty ship, escort carrier) were much more effective at turning the tide of battle than designs that were on the cutting edge of technology (eg. the Vought F4U Corsair, Gloster Meteor, Japanese Shinano aircraft carrier). The latter were often better in absolute performance, but arrived late, in small numbers, and with teething troubles that made the former carry the bulk of the battle. The Liberty Ship, for example, used reciprocating steam engines that were 50-year-old technology in WW2, but they were "good enough" and dead simple to make.
Combine that with the US's ability to unilaterally destroy Taiwan's fabs, and it sways the calculation a bit
I mean, they might if Intel were allowed to fail.
Intel is in the midst of a dramatic turnaround and huge shift in strategy. It might fail. But if they succeed it puts Intel and the US in a much stronger position in terms of technology and military leadership.
The end result is more like all the rich people take their cash and jump off the top of the pyramid as it crumbles
It will take decades for the US to get where Taiwan is now in semiconductor manufacturing, if ever. It's not just about building the most advanced chip factory. It's about re-aligning the entire nation's value system and culture to allow such development to happen in the first place.
We complain about the money we spend already. And now we're supposed to subsidize an entire industry to the point where we can build the most complex machines known to civilization at scale in a time-frame that matters to a global conflict that's potentially approaching soon? I don't see it.
https://www.tomshardware.com/tech-industry/us-govt-pushes-nv...
Hygon still seems to be making x86 CPUs: https://www.techpowerup.com/336529/hygon-prepares-128-core-5....
The others would probably be GlobalFoundries, Micron, Microchip, and TI.
Intel is:
* Critical to national security
* An advanced, industry that's extremely hard to spin up
* Essentially, one of two companies in it's industry.
Very few other companies meet all of those criteria.
TSMC aims to have N3 in Arizona by 2028 at the earliest which is 6 years after it first released. By that time, TSMC will have released N3X, N2, N2P, N2X, A16, and A14.
TSMC is heavily sponsored by the Taiwanese government and was created with the express purpose of making Taiwan so valuable that the West would be forced to defend them against China. Moving newer processes out of the country is against their national interests and they've made it clear that there's no plan to do that.
You’re reading a lot into the US right now. US policy in 2025 is more about which member of the whack pack is the alpha gorilla than anything else.
Future wars are likely going to be GPU driven, ML heavy entities where efficiency matters a lot more than brute force, blunt grenade throwing wars of the past.
A super power like US would likely want to be in the forefront of this if they happen to be in a tussle with a worthy adversary.
Imagine a next-generation fire-and-forget weapons with radar and broad-spectrum camera arrays and an AI trained on a fused version of all this data. Typical defenses like chaff or flares would be rendered almost entirely useless.
This kind of visual approach also renders modern stealth almost completely useless. When an unexpected plane is found on L-band (or some other low-frequency radar), the AD would simply fire a couple of missiles into the area with instructions to visually identify the large objects moving at fast speeds using the fusion of these different sensors (and ground+air-based radars) while in flight.
We are getting pretty close to being able to do this in realtime with cellphone-level chips.
(like the failed Clipper chip) https://en.m.wikipedia.org/wiki/Clipper_chip
The thinking might be the government needs a local industry for security. Think submarine manufacturing. Not a huge private market for that, but best to keep local so the supply can’t be cut off.
Though usually the government isn’t the best stewards of companies. When I worked for a large government contractor someone joked “yesterday’s technology tomorrow”. Some of that is for reliability, but it wasn’t cutting edge in a lot of ways.
"Giveaway?" This isn't some secret, everyone knows the military depends on x86 processors, and having a company that can produce them domestically is a national security concern.
Can you point out which specific findings? Ideally ones that are substantiated and not just one off tweets.
It's taken about 8 years to realign the US from a democracy to a fascist regime, something that was nearly unthinkable. This isn't a hard problem with the right propaganda and manipulation.
Switzerland owns its energy companies and its public transport company. Hugely successful.
Lookup Elbridge Colby - served in first Trump admin and now Undersecretary of Defense. Along with Hegseth and JD Vance, they are all that same line of thought
China would not takeover TSMC intact. Even if they did, they would not be able to operate it for quite some time (years), if ever.
https://www.youtube.com/watch?v=LMRYvl2Jefg
Hegseth on China
https://www.defense.gov/News/News-Stories/Article/Article/41... https://edition.cnn.com/2025/03/27/asia/pete-hegseth-asia-to...
I don't think the views of all the "high level strategist" types align precisely. But there are very few true isolationists like Rand Paul in the executive branch
I find it doubtful that the current admin would just let China walk in to Taiwan. Trump doesn't want a war, but he's not going to want to get bent over and make the US look weak either.
And its not all that different from Democratic position, its a bit of "Washington Consensus" type of situation, like anti-Communism was during the Cold War. The approach between admins is slightly different however, and the Trump admin doesn't like Europe all that much.
Fabs run on data. It takes years to gather that data.
Fabs can't just be repurposed overnight. Yields must be good, which takes data, equipment, etc which is all extremely specialized.
Very ignorant take.