I never thought this would happen, or that if it did, we'd be a few generations behind.
Now let's onshore or friendshore everything else we need! Rare earths, mid-tier processors, chemical precursors, pharmaceuticals, steel, robotics/mechatronics, solar, drones, ...
Why even stop there? Kill the Jones Act, get back to building naval drones and ships of all kinds, ...
NVIDIA has often lagged on process, since they drive such large dies, but having the first major project demo wafer on N4 now is literally 2 generations behind Taiwan.
When was the last time current gen, competitive GPUs were fabbed outside Asia?
US salaries are astronomically high compared to the rest of the world. In the tech sector that's doubly so. Everything is incredibly expensive there. Is this basically a small facility to keep some politicians happy?
Or is it used to provide some supply military gear at 50x the price?
Will it get shut down in a few years once everyone forgets about it?
Agree except this point, as at least for MCUs which are of US origin and yet cost ~$1 per chip which also includes reel packaging and distribution margins etc.
Getting the factory built, however, takes significantly more time and costs more in both actual dollars but, more importantly, the opportunity cost lost to not producing products in that extra build time.
I'd wager combining that with US defense contracts for US made chips would be lucrative for NVidia.
There's going to be a bunch of people wanting any job [1] and more in the coming years. So they'd probably cheer a local walmart as much as a local tsmc.
> TSMC is known for overpaying
Never heard that in my life. Underpay sure.
They're kind of notorious for have overqualified workers. PhDs left monitoring assembly lines or writing accounting software
My bad - didn't see the qualifier for Taiwan there.
Yea, TSMC pay is decent by Taiwan standards, but imo that's largely because Taiwan salaries are fairly low even compared to other East Asian nations when factoring CoL.
No doubt keeping costs low helped them get to that stage. Still that's a crazy margin. I imagine much of it's re-invested.
IMHO, Intel could've kept up but dropped the ball pretty hard. The volume of ARM / cell phone cpus caught them off guard and huge amounts of revunue let TSMC drive forward.
And if US could somehow provide lower electricity cost to Fabs, which is the number one cost in production, it will offset a lot of expensive items on the list. The actual labour cost for running the Fab is comparatively small.
Thus, salaries and cost of services do not factor in as heavily as you might think to fab economics.
Data suggests that TSMC's per wafer costs in Arizona are 10-30% higher than Taiwan and that Arizona fab is relatively new. It's economics will probably improve over time, narrowing the margin to 5-15%.
Looking towards the future, power costs and other global supply chain factors could very easily make TSMC's Arizona fab less expensive and more reliable to operate over time. For one, the US is completely energy independent... Taiwan is not.
I'm not saying tiny lines aren't cool. I'm just saying the idea you can't be successful if you make cruder, older resolution chips is probably wrong: Your printer and your car don't care if the Dice is 10mm not 5mm, and the track lines are 5x wider. MILSPEC stuff probably runs cruder for other reasons. Resiliency? Verilog proofs?
I also have no idea how many dice you get off a single ingot these days. 300mm wide, but how long?
Yeah but much larger(16-12nm) and much less profitable nodes than what Taiwan, the US or even Japan and China have now.
> I'm just saying the idea you can't be successful if you make cruder, older resolution chips is probably wrong
Define success. Smallest nodes are bringing in the most profits and every country prefers more profits versus less profits, especially Europe given it's budget deficits and welfare spending.
Larger nodes that aren't very profitable are good for national security but Russia and even North Korea are proof you don't need much domestic semiconductor industry to completely terrorize neighboring countries and level entire cities. WW1-style artillery shells and rifle rounds will do just fine.
I don't buy labour costs. I think it's probably post-VLSI packaging, assembly of devices, and compliance with HAZMAT that made the moves to Asia and Latin America happen. These plants don't use a lot of cheap labour. So wage cost cannot explain the decision.
There are interesting posts going back decades (crufty google doc shares, USENET posts..) talking about what the work culture at TSMC is like, about Intel management. I think VLSI is like cheese making: if you wear the wrong perfume on the wrong day, you destroy the entire output in one go. But that doesn't justify 16 hour days and toxic management.
roboror•3mo ago