There is a specific reason the American effort to de-risk its pharmaceutical supply chain keeps producing announcements without producing independence. The policy is aimed at one layer of the manufacturing stack, and the chokepoint sits two layers below it.

Making a generic drug is a vertical process. Key starting materials, basic industrial chemicals, get converted into intermediates, intermediates get converted into the active pharmaceutical ingredient, and the API gets formulated into a finished pill. Public attention and most policy attention land on the API, because that is the recognizable ingredient with the drug's name attached. But China's grip is tightest at the bottom of that pyramid, at the key starting materials and intermediates, where high environmental costs and low margins have driven Western and even Indian producers out of the market.

That distinction is the whole problem. A reshored API plant that buys its starting chemicals from China has relocated the visible step and left the actual dependency untouched. It looks like resilience on a map and behaves like dependency under stress.

Why India is not the alternative

The same structure explains why the most commonly proposed fix does not deliver what it promises.

India is routinely described as the pharmacy of the world and the natural second source, and it does have genuine volumetric capacity. It is also, itself, a customer. Roughly 70% of India's API needs, and up to 90% for critical antibiotics like cephalosporins and penicillin, are met by Chinese imports, and Indian government data through fiscal 2024-25 showed that pattern barely moving.

So a US buyer who shifts finished-dose or API sourcing from China to India has diversified the last step of a chain whose earlier steps run through the same place. That is not worthless, since a second finished-goods supplier does add flexibility against a factory fire or a single-plant inspection failure. But against the scenario that actually motivates the policy, an export restriction at the source, it provides much less protection than the org chart suggests.

The procurement system is the root cause

Here is the mechanism that built the dependency, and it is domestic rather than geopolitical.

Consolidation among pharmaceutical wholesalers produced a generic market of ruthless price competition, in which wholesalers solicit daily tenders and choose the lowest price on offer, then compete for hospitals and pharmacy chains on price rather than quality, availability, or manufacturing resiliency.

Sit with what that does to a manufacturer's incentives. If your buyer runs a daily auction and awards volume to the lowest bid, then every dollar spent on redundant capacity, domestic sourcing, larger inventory buffers, or a second supplier is a dollar that makes you lose the auction. Resilience is a cost, and the purchasing system was built to punish costs.

This is why reshoring campaigns keep underdelivering. Build an American KSM plant with American environmental compliance and American labor, and it produces a more expensive chemical that loses the daily tender to a Chinese competitor. Any serious resilience policy therefore has to include a buyer willing to pay for resilience. That could be government procurement with domestic-sourcing requirements, long-term contracts that guarantee volume rather than daily bids, or standards that require purchasers to weigh supply security alongside price. Whatever the instrument, the constraint is the same: someone has to pay more, and so far the system is designed so that nobody does.

The tariff results so far

The main tool actually deployed has been tariffs, and the empirical record is worth stating plainly. Despite years of tariff pressure, China has strengthened its position as the dominant API supplier, with hundreds of FDA-registered API facilities, and US exposure runs as high as 80% for certain ingredients.

There is a plausible explanation for why the tool did not produce the intended result, and it follows from the procurement point. Tariffs raise input costs for the generic manufacturers buying those ingredients, and those manufacturers already operate on margins thin enough that several have exited the market entirely. Raising their costs without changing what buyers will pay squeezes the weakest domestic players first, which can accelerate consolidation and exit rather than reshoring.

The measurement problem underneath

There is a further complication that makes all of this harder to manage: the United States cannot fully see its own supply chain. A 2024 Government Accountability Office assessment concluded the country lacks visibility into foreign manufacturing for 63% of critical drugs, with sourcing often running through intermediaries that obscure quality and origin.

That is a governance problem before it is a geopolitical one. A country that does not know which of its critical medicines depend on which foreign plants cannot prioritize, cannot stockpile intelligently, and cannot tell whether a given reshoring investment addresses a real chokepoint or a symbolic one.

What is actually being tried

Several governments are moving, and the approaches differ in ways worth noting.

The US has directed the Administration for Strategic Preparedness and Response to build a strategic reserve of active ingredients, identifying roughly 26 critical medicines and targeting a six-month supply of domestically manufactured APIs where possible. A stockpile is a genuine hedge against a sudden cutoff, and it is a bridge rather than a solution, since six months buys time to respond but does not create production capacity.

The European Union's Critical Medicines Act has expedited approvals for manufacturing medicines dependent on Chinese supply chains. Japan is restarting and upgrading domestic production of critical antibiotics made by fermentation, which is notable because fermentation capacity for antibiotics is one of the hardest things to rebuild and one of the most concentrated. India's incentive schemes target starting materials directly.

The efforts aimed at the bottom of the pyramid, Japan's fermentation restart and India's KSM incentives, are better matched to the actual chokepoint than the ones aimed at final assembly. That is the useful test to apply to any new announcement: which layer does it operate on, and does the input for that layer still come from the place being de-risked.

Reading it honestly

The security concern is legitimate and does not depend on any particular view of China. Export restrictions on medical goods during the early pandemic were not hypothetical, and a country that cannot make its own antibiotics has a real vulnerability regardless of who its trading partners are.

At the same time, the diagnosis in most of the public debate is aimed a layer too high, and the remedy most often proposed, tariffs plus reshoring announcements, addresses geography while leaving the economics that produced the geography intact. China's position was not primarily won through strategy. It was won because Chinese producers were willing to make low-margin chemicals with environmental costs that Western producers were not, for buyers who selected on price alone and asked no other questions.

Reversing that requires accepting the inverse: paying more for medicines that are more expensive to make, on purpose, indefinitely. That is a defensible policy choice and an unpopular one, and it is the actual decision underneath the supply chain debate. Until it is made explicitly, the plants will keep being announced, the ingredient of record will keep moving closer to home, and the chemicals those plants start from will keep arriving from the same place they always did.

Further reading