The United States committed hundreds of billions of dollars to bring chip manufacturing home. The fabs are rising in Arizona, Ohio, New York, and Texas. The problem now is not concrete or lithography tools. It is the technicians who keep a fab running, and the country is not producing enough of them.
How big is the semiconductor workforce gap?
The Semiconductor Industry Association and Oxford Economics project the U.S. semiconductor workforce will grow by about 115,000 jobs by 2030, from roughly 345,000 to 460,000. Of those new positions, about 67,000 risk going unfilled, close to 58 percent. And the gap has a shape:
| Segment of the ~67,000 gap | Share | Roughly |
|---|---|---|
| Technicians (certificate / 2-year) | ~39% | ~26,000 |
| Engineers (4-year+) | ~41% | ~27,000 |
| Computer scientists | ~20% | ~13,000 |
Source: SIA and Oxford Economics, "Chipping Away," 2023. Technicians are the single largest job category at risk, and the one a trade pipeline can fill fastest.
Which fabs, and how much is committed?
| Fab | Investment | Direct jobs |
|---|---|---|
| TSMC: Arizona | ~$165B (6 fabs) | ~12,000 |
| Micron: New York | up to $100B | ~9,000 |
| Samsung: Taylor, TX | $37B+ | ~3,500 |
| Texas Instruments: Sherman, TX | up to $30B (4 fabs) | up to 3,000 |
| Intel: Ohio | $28B+ | ~3,000 |
Company announcements; figures are announced totals and job estimates, not all yet operational. Several projects have slipped on schedule for market and capital reasons.
Why technicians specifically?
Because they are the roles a trade program can produce, and the ones the industry stopped investing in. A fab technician usually holds a certificate or a two-year degree, often from a community college, and installs, runs, and maintains the equipment that makes chips. For decades the industry funneled its training dollars toward degreed engineers and research, and the technician pipeline thinned. Now about a third of the current semiconductor workforce is nearing retirement, so the fabs are hiring into a shortage that was building long before the CHIPS Act.
Has the shortage already caused delays?
Yes, and publicly. TSMC cited an insufficient number of skilled workers for equipment installation when it pushed mass production at its first Arizona fab from 2024 to 2025, then flew in more than 500 experienced technicians from Taiwan to bridge the gap. That is the clearest documented case of a workforce constraint moving a U.S. fab's schedule. This is not a Coherent-only story either. Arklight's Coherent Sherman fab brief shows the same pattern at the photonics layer, in the same Texoma labor pool that TI is drawing on.
What is at stake?
The whole point of the buildout. The United States is on track to roughly triple its semiconductor manufacturing capacity by 2032 and to go from making zero percent of the world's most advanced logic chips to about 28 percent. That reshoring is a national-security goal. The technician gap is the single thing that most directly threatens it, because a fab with no one to run it produces nothing, no matter how much capital built it.
Who trains them?
Community colleges and fab apprenticeships are standing up, but throughput has to beat both new demand and retirements at once. Closing that gap means producing production-ready technicians faster than the old pipeline can, which is the problem the Arklight Demand Model measures and that Trade School 2.0 is built to solve. For the employer's version of the same question, see manufacturing workforce development.
The bottom line
America bought the fabs. It has not built the workforce to run them, and technicians are where the gap is widest and easiest to fill. Reshoring the chip industry is a training problem now, and the fabs cannot wait five years for the answer.