During our recent router discussion, the topic of computer chips came up. While a shift in policy has the potential to disrupt the availability of consumer-grade routers, the bigger implication, we found, was the continued short supply of chips—which might only get worse with the expanding footprint of AI data centers. That thought led me to wonder: What ever became of all the funding and foundries that were supposed to stand up after the passage of the CHIPS and Science Act back in 2022?
So, I did a little digging.
What I found (which is laid out in more detail below) is that there’s been a significant amount of award dollars distributed and projects announced and started in the time since the bill was enacted, both private and public. But the intended impact—the influx of new U.S.-made chips—is still potentially years away. That said, when we get there, the impact can, and likely will, be significant. It’s just going to take a little longer.
As of January of 2025, the U.S. Commerce Department’s Office of the Inspector General reported 20 incentive-based awards totaling about $33.7 billion in direct funding and $5.5 billion in loans related to the CHIPS Act. In addition, the CHIPS Research and Development Office has awarded some $8.3 billion to operate the National Semiconductor Technology Center and to fund research projects that will bring new CHIPS technology to the commercial market. That’s a lot of dollars distributed, but, again, the payoff for those investments is still years away.
On the private side, the story is very similar. There are projects in the pipeline that will bring meaningful change—once completed. According to the Semiconductor Industry Association (SIA), there have been more than 140 projects across 30 states that amount to over $640 billion in private investment since 2020. Those projects will help create and support over 500,000 jobs (70,000 facility jobs, 122,000 construction jobs, and over 335,000 additional jobs).
But what about actual supply relief right now?
TSMC’s Arizona fabrication facility (which will eventually include three fabs) is the clearest sign of progress on this front. Their Fab 1 facility, which produces 4nm and 5nm FinFET process technologies, ramped up to full-volume production in the first half of 2025, becoming the first “success story” of the CHIPS Act. TSMC received $6.6 billion in direct funding from the program and a $5 billion loan that went towards their $65 billion in capital expenditures for the project. Once complete around 2030, the three-fab facility will have created 20,000 construction jobs and “tens of thousands of indirect jobs.”
Looking at the other projects in the pipeline as shared by SIA, there are roughly 18 new fab construction projects ongoing throughout the country with many of them expected to be operational between 2027 and 2028. The majority of projects, though, look to be on track for completion around the end of the decade, with production ramping up in the early 2030s.
Meaningful yields, then, likely won’t be felt until the mid-2030s.
When they are realized, though, they will be significant. According to SIA, the U.S. will have tripled its semiconductor manufacturing capacity by 2032, increasing its share of global fab capacity from 10 percent to 14 percent, and its share of advanced sub-10nm manufacturing to 28 percent. SIA said that without the CHIPS Act, the U.S. would reach around 9 percent of global semiconductor capex, compared to the 28 percent under the current trajectory.
On the leading-edge memory side (the area where AI is gobbling up supply), the U.S. on pace to capture around 10 percent of global leading-edge DRAM production by 2035, which is up from 0 percent today. That is mainly influenced by ongoing Micron projects.
TL;DR: The CHIPS Act is working, but actual chip supply chain impact won’t be felt for a few more years.
