Loading issue research
Continue federal chip-fabrication grants and tax credits, reject government equity stakes, and fix the H-1B rules undercutting the workforce these fabs need.
Verification Status
AI-researched, unverifiedLast Reviewed
Jul 4, 2026
Cited Sources
7
What is failing, what we would change, and the conclusion we are willing to defend.
Domestic semiconductor manufacturing is one of the bipartisan success stories in recent industrial policy. It's also, as of 2025-2026, a live test case for how not to handle the government's role in funding it. The 2022 CHIPS Act passed with bipartisan support on national-security grounds. By 2026, that bet is paying off. TSMC's Arizona fab is profitable and running ahead of schedule. Intel has crossed a threshold no US fab had reached before, 18-angstrom-class production, live in Chandler. Micron and Samsung are underway too, unevenly: Intel's Ohio site has slipped years, Samsung's Texas site has been delayed. The wager that domestic fabrication could work here again isn't hypothetical anymore.
But 2025 introduced a new and, this issue argues, mistaken mechanism: converting Intel's unpaid CHIPS grants into a roughly 10% federal equity stake in the company. That drew objections from free-market conservatives and consumer advocates on the left for different but converging reasons: dilutive terms for existing shareholders, no shareholder vote, and a precedent where the government becomes a company's part-owner. Conditional grants already provide leverage without ownership. Separately, a new $100,000 fee on new H-1B visa petitions raises the direct cost of exactly the skilled-engineer pipeline (Taiwanese and other foreign process engineers) that TSMC and Samsung rely on to staff and start up these fabs. That's an immigration policy working against the administration's own onshoring goal.
There's a fourth issue this platform won't pretend doesn't exist: leading-edge fabrication uses enormous volumes of ultra-pure water, and it's being built out fastest in Arizona, a state already absorbing Colorado River cuts. The honest picture is more nuanced than either "chip fabs are draining the desert dry" or "there's nothing to see here": substantial consumption, substantial reclamation investment, and a still-unresolved history of semiconductor manufacturing contaminating groundwater the last time this industry scaled up this fast, in the original Silicon Valley.
Continue direct federal support for domestic chip fabrication, which is not a partisan question and shouldn't become one, but structure it as conditional grants and tax credits with public accountability (reporting, workforce, and delivery conditions), not government equity stakes in the recipient company, regardless of which administration proposes the equity model.
Carve out an explicit exception (or a fast, predictable national-interest waiver process) from the $100,000 H-1B fee for the specific, narrow category of skilled semiconductor process and equipment engineers needed to bring a CHIPS-funded fab online. The fee currently taxes the same onshoring goal it's nominally unrelated to.
Expand federal semiconductor workforce-training funding to meaningfully close the projected ~67,000-worker gap by 2030, and pair any company's CHIPS funding with enforcement of the non-discriminatory domestic-hiring commitments made to get that funding in the first place.
Require CHIPS-funded fabs to publish independently audited (not self-reported) water-use and recycling figures, and require environmental-remediation bonding sufficient to cover a large-scale groundwater contamination event before a fab receives federal funding. This proposal learns directly from the fact that the last time this industry scaled up this fast, in the original Silicon Valley, it left behind one of the country's densest concentrations of EPA Superfund sites.
Continue federal chip-fabrication grants and tax credits, reject government equity stakes, and fix the H-1B rules undercutting the workforce these fabs need.
Turn frustration into useful pressure.
If this position misses evidence or a lived consequence, challenge it. If it holds up, help test it locally and connect it to the issues around it.