The CHIPS Act Is Working. Now Watch Your Supply Chain Break Anyway.

$39 billion in CHIPS Act subsidies have funded 23 semiconductor fabrication projects across the United States as of May 2026. Intel, Samsung, TSMC, Micron, and a dozen other players are pouring concrete for fabs in Arizona, Ohio, Texas, and upstate New York. The optics are flawless. The national security narrative is locked. The bipartisan political victory is secure. And if you run an industrial operation that depends on semiconductor supply, you should be very worried about what comes next.

Here is what the math actually says: the incentive structure has created winners and losers, and the loser is you.

The CHIPS Act was engineered to answer a real problem. In 2022, semiconductor shortage was hemorrhaging American manufacturing. Lead times stretched to 18 months. Automotive plants shut down. Medical device makers rationed inventory. The policy solution was logical: subsidize domestic fab capacity so the United States stops depending on Taiwan and South Korea for advanced chips. Build the fabs here. Shorten supply chains. Reduce geopolitical risk.

The implementation is where the framework breaks. The subsidies are structured as direct funding to fab operators, not to the manufacturers who will actually use the chips. TSMC, Intel, and Samsung get the cash. They build the fabs. They own the capacity. They control pricing. And they have no contractual obligation to sell chips to you at a discount just because American taxpayers funded their expansion.

The first sign appeared in late 2024. Intel announced its Arizona fab would take orders, but qualification would take 18 months and minimum order quantities would be high. Samsung's Texas fab ran into the same problem: qualified customers only, lead times measured in quarters, volume commitments required upfront. TSMC's Arizona operation follows the same playbook. These are world-class manufacturing operations. They are also closed ecosystems that serve large, established customers. Not your supplier base.

The second problem is geographic. The CHIPS Act fabs are clustered in four states: Arizona, Ohio, Texas, New York. That sounds distributed until you look at logistics. Arizona is a desert with one major interstate. Ohio has rail but limited ports. New York is far from both coasts. Texas has Houston, but fab construction is centered 200 miles inland. A semiconductor fabrication plant generates zero economies of scale if the chips take 72 hours to reach the customer. You save two weeks of ocean freight from Taiwan and gain three weeks of trucking through the Midwest. The supply chain math does not actually improve.

Third: the fabs are overbuilt for what they will produce. A modern advanced fab costs $20 billion to $30 billion and runs at full capacity or it hemorrhages money. TSMC's Arizona fab was built to serve Apple, which needs millions of chips quarterly. Intel's Ohio operation was designed around the automotive industry, which buys in volume. If your company needs 50,000 specialty microcontrollers or edge processing chips, you are not the target customer. You are the inventory risk that gets deprioritized when a major customer needs to scale.

The evidence is already visible in procurement data from early 2025. Companies that tried to qualify for CHIPS Act fab capacity found the process takes 12 to 18 months. Design review, sample validation, process qualification, risk assessment. It is the same timeline as sourcing from Taiwan, minus the geopolitical premium. One equipment manufacturer we spoke to abandoned the effort after six months and renewed their TSMC contract at higher pricing, simply because they knew the lead time. Another pivoted to legacy node fabs in Arizona and New Mexico, which are cheaper but do not solve the advanced semiconductor shortage.

The fourth problem, and the one nobody is discussing, is that the CHIPS Act created a two-tier industrial base. Large OEMs with sophisticated supply chain operations can navigate the qualification process and secure CHIPS Act fab capacity. Everyone else gets pushed back to the spot market and legacy suppliers. If you are a mid-sized industrial automation company, a precision instrument maker, or a specialty equipment builder, the CHIPS Act does not help you. It makes things worse because it drains available capacity from the suppliers you rely on. TSMC used to take smaller orders because American demand was fragmented. Now they consolidate customers into CHIPS Act programs and ignore the rest.

So what should you do? First, stop assuming reshoring automatically solves supply chain risk. It does not. Reshoring changes the shape of the risk. Second, do not wait for CHIPS Act capacity to solve your semiconductor sourcing problems. It will not arrive in time, and when it does, you will not qualify for it. Third, start building redundancy in your supply base now. If you currently source a critical component from one supplier, add a second source from a different geography, even at higher cost. The resilience premium is cheaper than a production shutdown.

The CHIPS Act is working exactly as designed: it is building American semiconductor capacity and reducing dependence on Taiwan. It is just not working for you. That distinction matters when you are trying to hit your production targets next quarter.