Ohio One Construction Timeline Update
newsroom.intel.com
newsroom.intel.com
> The semiconductor manufacturer originally estimated its first factory would come online in 2025. Then, it bumped the grand opening to 2027. Now, Intel estimates operations will begin at the Ohio One campus somewhere between 2030 and 2031.
The biggest issue is that there isn't sufficient demand for an Intel fab, which is causing them to slow play opening it. Ben Thompson has written at length on Intel's struggles over the years and advocates that the U.S. step in and help guarantee sufficient volume of orders from US customers to make the fab financially work:
> That leaves Intel and the need for native leading edge capacity, and this is in some respects the hardest problem to solve.
> First, the U.S. should engineer a spin-off of Intel’s x86 chip business to Broadcom or Qualcomm at a nominal price; the real cost for the recipient company will be guaranteed orders for not just Intel chips but also a large portion of their existing chips for Intel Foundry. This will provide the foundational customer to get Intel Foundry off the ground.
> Second, the U.S. should offer to subsidize Nvidia chips made at Intel Foundry. Yes, this is an offer worth billions of dollars, but it is the shortest, fastest route to ground the U.S. AI industry in U.S. fabs.
> Third, if Nvidia declines — and they probably will, given the risks entailed in a foundry change — then the U.S. should make a massive order for Intel Gaudi AI accelerators, build data centers to house them, and make them freely available to companies and startups who want to build their own AI models, with the caveat that everything is open source.
> Fourth, the U.S. should heavily subsidize chip startups to build at Intel Foundry, with the caveat that all of the resultant IP that is developed to actually build chips — the basic building blocks, that are separate from the “secret sauce” of the chip itself — is open-sourced.
> Fifth, the U.S. should indemnify every model created on U.S.-manufactured chips against any copyright violations, with the caveat that the data used to train the model must be made freely available.
The shift from 2027 to 2030 is the first material slippage -- and reflects one or two things (or both!): 14A process technology timeline and/or soft foundry demand from the fabless firms.
This is so absurd it's getting to the point where it's laughably evil.
No, the US has far simpler and completely free methods to do this. For example, look at how travel works. You can't use any US federal dollars to fly on non-US airlines unless it's impossible to do so (with some other caveats we won't go into here).
What the US should do is simply declare, no US federal dollar can go to any AI chip that's not produced in the US. No US federal dollar can go into running training or inference for any model that isn't using GPUs produced in the US.
That's it. An instant massive market is created. Without taxpayer dollars. Without subsidies.
But we can either make things cost a little more or give away massive piles of money while getting maybe nothing in return and have everything cost more anyway.
Have those funds all been dispersed?
It’s spent over 50B.
One of these things is not like the others.
Isn't that a massive giveaway to a relatively small number of model makers, at the cost of the huge number of copyright holders, including international copyright agreements?
If this was legislation, I would assume that someone bribed and/or horse-traded for it to be tacked on.
Copyright would be the single pillar of the international rules-based order to survive this dumpster fire.
But this plan from Ben Thompson is laughable.