What the CHIPS Act Actually Funds
The CHIPS and Science Act, signed into law in August 2022, allocates $52.7 billion for US semiconductor manufacturing and research. The headline number gets thrown around a lot, but the breakdown matters:
- $39 billion in manufacturing incentives — direct subsidies to companies building or expanding fabs in the US
- $11 billion for R&D programs, including the National Semiconductor Technology Center (NSTC) and the National Advanced Packaging Manufacturing Program (NAPMP)
- $2 billion for mature node and current-generation chip production (defense and critical infrastructure supply chains)
- $500 million for international semiconductor supply chain coordination
- Plus a 25% investment tax credit for semiconductor manufacturing equipment (scored separately from the $52.7 billion)
The $39 billion in manufacturing incentives is the part driving actual fab construction decisions. Companies apply to the Commerce Department's CHIPS Program Office, which evaluates proposals based on technology advancement, job creation, community investment, and clawback provisions if the company doesn't meet its commitments.
The New Fab Environment
As of mid-2026, the major CHIPS Act-funded projects include:
TSMC in Phoenix, Arizona — three fabs planned, with a total investment exceeding $65 billion. Fab 1 (N4 process) began limited production in late 2025. Fab 2 (N3/N2) targets 2028. Fab 3 (N2 and beyond) is planned for early 2030s. CHIPS Act funding: $6.6 billion plus up to $5 billion in loans.
Intel in Chandler, Arizona and New Albany, Ohio — four new fabs total. Arizona expansion adds two fabs to the existing Ocotillo campus. Ohio is a greenfield site with two fabs. Combined investment: over $100 billion across both sites. CHIPS Act funding: $8.5 billion plus $11 billion in loans. Intel is also expanding in Leixlip, Ireland and Magdeburg, Germany (funded by EU CHIPS Act equivalents).
For a related perspective, see Yield Engineering: Why Chip Manufacturing Yield Matters More.
Samsung in Taylor, Texas — one fab initially, with plans for additional facilities. Total investment: $17+ billion. CHIPS Act funding: $6.4 billion. The process target is Samsung's advanced gate-all-around technology.
Micron in Clay, New York and Boise, Idaho — DRAM manufacturing complex in New York (largest private investment in state history) plus expansion in Boise. Total investment: up to $100 billion over 20+ years. CHIPS Act funding: $6.1 billion. The NY fab will produce leading-edge DRAM.
Smaller grants have gone to GlobalFoundries ($1.5 billion for expansion in Malta, NY and Burlington, VT), BAE Systems ($35 million for rad-hard chip production), and various packaging and mature-node facilities.
The Economics of Onshore Manufacturing
Building chips in the US costs more than building them in Asia. How much more is a matter of debate, but credible estimates range from 30-50% higher construction costs and 20-40% higher operating costs compared to Taiwan or South Korea. The differences come from:
This connects to the ideas in HBM Packaging: Through-Silicon Vias, Microbumps, and Why HBM.
- Higher construction labor costs — a fab construction worker in Arizona earns 3-4× what one in Taiwan does
- Higher operating labor costs — technicians and engineers cost more, and the US talent pool is smaller (see the talent crisis)
- Permitting and regulatory compliance — environmental reviews, building permits, and utility connections take longer in the US
- Supply chain distance — many specialty chemicals, gases, and sub-components for semiconductor equipment are sourced from Asia. Shipping them to Arizona or Ohio adds cost and lead time.
- Utility costs — electricity for a leading-edge fab runs $100-200 million per year. US industrial electricity rates vary widely but are generally higher than Taiwan's subsidized rates.
The CHIPS Act subsidies offset some but not all of this cost differential. A $6 billion subsidy on a $40 billion fab project covers 15% of the capital cost. It helps — it can be the difference between a project's financial model working or not — but it doesn't eliminate the fundamental cost advantage of Asian manufacturing.
The Real Motivation: Supply Chain Risk
If pure economics drove the decision, no one would build leading-edge fabs in the US. The real driver is geopolitical risk. About 92% of advanced chip manufacturing (sub-10nm) happens in Taiwan. A Chinese military blockade or invasion of Taiwan — however unlikely — would be catastrophic for the global economy. The 2020-2022 chip shortage, which was caused by mere demand spikes and logistics disruptions, cost automakers alone an estimated $210 billion in lost revenue.
The CHIPS Act is essentially an insurance policy. The premium is high — tens of billions of dollars in subsidies — but the insured event (loss of Taiwan manufacturing) would cause trillions in economic damage. Framing it as insurance rather than industrial policy helps explain why bipartisan support exists despite the enormous cost.
Early Results and Challenges
The implementation hasn't been smooth. TSMC's Arizona fab was delayed by about a year due to construction challenges and the difficulty of finding qualified workers. The company brought in hundreds of Taiwanese workers to supplement the local workforce, which caused tension with American labor unions and employees who felt sidelined.
See also: Chip 2nm là gì? Tổng quan Công nghệ Bán dẫn Thế hệ Mới 2026.
Intel's financial difficulties have complicated their ambitious fab expansion plans. The company announced cost-cutting measures in 2024 that raised questions about the timeline for their Ohio project, though Intel leadership has repeatedly reaffirmed commitment to the CHIPS Act projects.
The Commerce Department's review process has also been slower than many in the industry expected. The first preliminary agreements weren't announced until early 2024, nearly two years after the law passed. Companies had to manage extensive application requirements including environmental reviews, community benefit plans, and commitments to provide childcare facilities for workers (a requirement that surprised some applicants).
Will It Actually Work?
By 2030, the US share of global semiconductor manufacturing is projected to increase from about 10% to 14-15% — meaningful but still far from the 37% share the US held in 1990. The CHIPS Act will create roughly 40,000-50,000 direct manufacturing jobs and perhaps 2-3× that in supporting roles.
The honest assessment: the CHIPS Act will establish meaningful advanced manufacturing capacity on US soil, providing a geopolitical hedge against Taiwan risk. It won't make the US self-sufficient in semiconductors — that would require sustained investment many times larger than the current act. And it won't change the fundamental cost equation that makes Asian manufacturing cheaper.
The real test comes in the 2030s, when the initial subsidy money runs out and these fabs need to compete on operational economics. If they can attract enough customer volume and achieve competitive yields, the ecosystem becomes self-sustaining. If not, the conversation about additional subsidies starts over.