When U.S. legislators proposed the CHIPS and Science act a couple of years ago, they set a rather vague goal to significantly increase domestic chip production by building new fabs. However, the U.S. government will set the wheels in motion next month by accepting applications for semiconductor projects. And this week, Commerce Secretary Gina Raimondo outlined the ambitious goals that the country has in mind with its CHIPS and Science initiative.
The Biden administration hopes that by 2030, the U.S. will develop and produce the world's most advanced chips on its shores. To bring the industry's most sophisticated process technologies to America, the U.S. government expects at least two new large-scale clusters of leading-edge logic fabs to emerge. Both should have a strong supplier ecosystem, specialized infrastructure, and R&D operations.
Also, the government expects chipmakers to deploy high-volume advanced chip packaging facilities in the U.S.
Both goals are achievable as Intel, TSMC, and Samsung Foundry are building advanced chip fabs in the U.S. Intel is also building up an all-new campus in Ohio that will include leading-edge fabs and advanced packaging facilities. While TSMC is expected to keep its most sophisticated nodes in Taiwan for the next few years, as its campus in Arizona expands (and it is rumored to include advanced packaging operations over time), it may start making chips on leading-edge nodes there as well. The same applies to Samsung Foundry's Taylor, Texas fab, which used to make its most advanced chips in Texas.
In addition to the world's most advanced logic and packaging facilities, the U.S. government wants the country to produce advanced memory chips 'on economically competitive terms, which Micron expects to achieve. It just so happens that Micron plans to deploy two new major memory fabs in the U.S. — in Idaho and Ohio — by the end of the decade.
Also, the American government wants chipmakers to expand production capacity for the current-generation and mature-node chips most critical to the country's economic and national security.
"I want the United States to be the only country in the world where every company capable of producing leading-edge chips will have a significant R&D and high-volume manufacturing presence," said Raimondo. "Now, achieving these goals won't be easy. We are ambitious, but we are not naïve."
Next week a special jury will begin to consider applications by various parties to receive funds and other incentives to build and develop chips in the USA.
$39 billion is allocated for manufacturing incentives, while another $11 billion will be invested to build a strong semiconductor R&D ecosystem.
Hopefully the government officials working on this act realize the parasitic relationship and plan for it. We need long-lasting chip fab production capabilities that are resilient to regional disruptions.
And wasn’t it their economic policies that drove the fabs away making them too expensive to operate in this country?
What we do know is that the trend lines were all going in the wrong direction, with semiconductor manufacturing leaving the US and going mostly to Asia. Furthermore, the industry has positively reacted to the prospect of CHIPS funding. So, from what I can tell, it so far looks to be working. We'll have to see how this all plays out, to know for sure.
Some amount of that will be unavoidable, if you want to maintain capacity in the US. The analogy with agriculture is apt - you have to keep farmers solvent both when their crops are wiped out and when bumper harvests cause prices to crash to unprofitable levels.
With semiconductors, I think two key principles should be:
The vast majority of funding should be private. Government funding should be used mainly to spur more risky investments and perhaps as loan-guarantees to de-risk private investment.
Funding should be tied to specific deliverables which would not be economically viable without it, and will not continue to receive government funding if/when cancelled.Still, there will likely come a point when the business cycle hits another trough and further injection of funding (i.e. CHIPS 2) is needed to avoid too much destruction of capacity.
That's an ideological narrative. Government can do big things. We used to believe this, like in the decades after WWII (examples: Panama Canal, Hoover Dam, Nukes, Interstate Highway system, Apollo Program, world's largest fleet of supercarriers). And we need only look at some of the amazing things China's government has accomplished, to believe it's still true.
No, I think the "pull" factors were much stronger than the "push" factors. In other words, China & other Asian countries provided strong incentives for foreign fabs to locate there. It's not just US fabs, either. Japanese and South Korean companies also built fabs in China.
The bigger issue isn't US manufacturers building abroad, so much as that other countries with strong semiconductor sectors enjoy a lot of state support, whereas the US has provided virtual none until now. In other words, it's a lot more what the US hasn't done that's caused the loss in competitiveness than what it has done.
That was their first in the US, no?