$20 Billion Grant For Intel
Bringing some chip manufacturing back to America.
Recently Additional Source:Asianometry had an important piece of the nature of chip foundries and what TSMC is doing to protect themselves from Chinese aggression.
Now Biden announces that we’re bringing chip manufacturing back to the U.S. with a $20 billion grant for Intel. What’s happening here is that Intel got behind TSMC due to their practice of incremental improvement. TSMC got the jump on them about fifteen years ago and has never let up. Intel will likely not catch TSMC, but they will become a source for things made only in America that will be used in defense AI applications.
This is good, this matches things Zeihan was predicting. Keep in mind Zeihan looks down from on high at this area, while Asianometry knows the gritty details, which we saw in From Game Graphics To AI.
Intel getting a grant like this is good, but Nvidia rolling out their new Blackwell Architecture is also a fine thing.
Powering a new era of computing, NVIDIA today announced that the NVIDIA Blackwell platform has arrived — enabling organizations everywhere to build and run real-time generative AI on trillion-parameter large language models at up to 25x less cost and energy consumption than its predecessor.
A 2500% increase in capability? What does this mean for the average GPU that might find its way here? I think that’s very much a datacenter thing, all the breathless speculation about the RTX 5090 is saying 70% improvement over the prior generation. I want to post a short video here about the new generation of processor but YouTube is full of crackheads who LARP as industry analysts. Rumor of a new GPU card has them all fulminating and I have zero patience for sorting through it all.
Conclusion:
We’ve been getting incremental slightly better every two years from Intel until TSMC lapped them. We’ve been getting a biannual boost from Nvidia, riding the TSMC wave. Competition is good, TSMC will move even quicker with Intel nipping at their heels. I’m excited to see a new generation of tensor cores, given how dramatically different it is shifting from my GTX 1060 to a virtual fourth generation device. I’m gonna hold my breath a little bit more and hope there’s a decline in the 4th gen pricing following the arrival of the new gear.
But I’ve been playing for time and I can only do that for just so long …