The first wave of CHIPS Act-funded fabs is moving from groundbreaking ceremonies to actual silicon production. TSMC’s Arizona Fab 21 is now producing 3nm chips at volume, Intel’s Ohio One fab is ramping 18A process technology, and Samsung’s Taylor, Texas facility is targeting advanced packaging alongside logic production. After years of political fanfare, the reshoring of semiconductor manufacturing is becoming a supply chain reality.
What’s Actually Online
TSMC Arizona Fab 21 Phase 1 entered risk production for N4P (4nm class) in late 2024 and reached meaningful volume in 2025. Phase 2, targeting N3 (3nm), is on track for 2026 volume production. Apple is confirmed as a Phase 1 customer, sourcing A-series and M-series chips for US-sold devices. The fab’s capacity is still a fraction of TSMC’s Taiwan operations β estimated at 20,000 wafer starts per month versus Taiwan’s 150,000+ β but it represents a meaningful hedge for US-based OEMs.
Intel’s situation is more complex. The company’s IDM 2.0 strategy β acting as both a chip designer and a foundry for external customers β has been challenged by continued process technology delays. However, 18A is now sampling with select customers, and Intel Foundry Services has signed several undisclosed fabless customers. The Ohio One fab is expected to reach initial production volumes in mid-2026.
What This Means for Buyers
For procurement teams, the practical impact of CHIPS Act investment is still mostly future-tense. The new domestic capacity won’t meaningfully affect commercial component availability for another 2β3 years. Most of the initial capacity is committed to strategic customers β defense primes, hyperscalers, and automotive OEMs β not the open market.
That said, there are already secondary effects worth tracking. The ramp of domestic capacity has created new demand for fab construction materials, specialized gases, and photolithography equipment β all of which feed back into supply chain constraints in adjacent categories. Companies like ASML, Applied Materials, and Lam Research are running full backlogs.
Geopolitical Hedge vs. Commercial Reality
The honest assessment for most commercial buyers: CHIPS Act fabs provide a geopolitical hedge against a Taiwan contingency scenario, but they won’t solve near-term allocation problems. The economics of US-based production are also significantly higher β TSMC Arizona wafers are estimated to cost 20β30% more than equivalent Taiwan production, costs that will eventually flow through to component pricing.
For supply chain strategists, the more actionable takeaway is geographic diversification of your supplier base β not necessarily US-only, but avoiding over-concentration in any single geography. Vyrian’s supplier network spans 40+ countries, giving buyers access to vetted sources across multiple manufacturing regions.
The Long Game
The semiconductor industry is playing a decade-long game of capacity investment. The fabs being announced today won’t reach full production until 2028β2030. Buyers who understand this timeline can plan accordingly β building supplier relationships now with fabs that will matter in the next decade, rather than reacting to the allocation crises of today.