The China panic is aimed at the wrong target

A biotech founder in Cambridge or San Francisco can now generate their drug's first human data in a trial run on the other side of the world for a fraction of what it would cost at home. In the first half of 2025, close to half of every new drug molecule that entered human trials anywhere in the world came out of a Chinese company, 46 percent by Goldman Sachs' count. With each passing month, China is looking more and more like the center of gravity of early drug development.

A trial in China can cost roughly 60 percent less and enroll patients several times faster than the same trial in the United States. And the thing being bought is no longer cheap routine labor. China now holds close to 30 percent of the global pipeline of innovative drug candidates, only a couple of points behind the United States (two years ago it trailed by thirteen). Its biotechs signed more than 130 billion dollars in out-licensing deals in 2025, well over double the year before. When results of that quality are available for a fifth of the price in less time, the work follows the gradient. A cash-strapped founder being asked by investors for human data does the arithmetic and books the trial where the arithmetic points.

We’ve noticed in recent conversations the loudest response has been to frame this as a question of national loyalty. Founders who develop abroad are somehow cast as taking American science and handing it to a competitor, even when that framing defies logic. No great founder can opt out of an economic gradient by force of patriotism, and asking them to is a way of avoiding the harder question underneath. The harder question is whether anything real is lost when this kind of work moves offshore at scale and, if so, what. 

Two concerns that deserve genuine weight come to mind, and neither has much to do with betrayal.

The first is about resilience. A country that develops and manufactures fewer of its own medicines is more exposed when something breaks, whether that’s a pandemic, a trade rupture, or a conflict that closes a border. This is the lesson the United States spent the last decade learning about semiconductors. The version of this concern that’s unwarranted (yet somehow widespread) is the fear that a drug tested abroad cannot reach American patients. This is simply not how approval works.

The serious version of this concern is different: that the capacity to originate and develop medicines, once it thins out at home, becomes something a country relies on others to supply. Today, something like four in five biotech companies depend on at least one Chinese partner somewhere in their research or manufacturing. This isn’t a crisis in itself – but it becomes one if the domestic alternative stops existing.

The second concern is more important, because it would make the first irreversible. First-in-human trials are the training ground for clinical operators, regulatory specialists, trial designers, and the CRO scientists who run the assays. When that work moves offshore, the domestic training ground shrinks with it. The Duke Clinical Research Institute has warned that once a country's early-phase muscle wastes away, the tacit knowledge inside it is slow and costly to rebuild. Ecosystems like this don’t decline in proportion. They have thresholds. Below a certain density of activity, the specialized suppliers and the experienced people stop being viable, and rebuilding then means re-crossing a threshold rather than topping up a reserve. This is what makes the erosion hard to see while it happens. No single decision to run a trial abroad causes it, and each such decision is individually rational, which is how a capability can slip away without anyone choosing to give it up.

To be fair, there’s a serious argument on the other side. Offloading expensive execution can free scarce domestic talent to concentrate on the highest-value work, the origination and design where the American edge has always been strongest. On this view, early-phase execution is closer to manufacturing than to invention, and a country loses little by letting it go where it is cheapest, the same way it loses little by offshoring the assembly of devices it still designs. What the ‘right’ view is depends on a question with no settled answer: whether early-phase development is routine work that can be safely handed off, or the soil in which the next generation of inventors is grown. Those most worried about China assume the latter. Those most comfortable with the shift assume the former. 

Regardless, both concerns above have the same shape. Both describe something that erodes slowly, driven by decisions that are each defensible on their own terms, and both grow expensive to reverse only once they have gone further than anyone intended. That shape doesn’t call for shaming founders for outsourcing work to China, which changes nothing. Nor does it call for walling the domestic industry off from competition, which would raise costs and slow the arrival of drugs people need. Competition in drug development is, on balance, good for patients, and cheaper, faster development is a gift rather than a threat. What the concerns call for is more prosaic and more difficult: keeping the domestic path worth choosing on its merits. 

That means capital willing to fund risk rather than only buy it once someone else has retired it, immigration choices that draw the best scientists in rather than turning them away, and above all speed. Australia clears a trial to begin within seventy days of a final protocol; in the United States the same study can wait years, most of the delay lodged in a pre-IND process everyone complains about and no one has fixed.

The worry we keep hearing is that American medicine will lose to Chinese medicine. That has never been the real danger. A world with more good drugs, developed faster and sold for less, is a better world wherever the work happens. The risk is subtler than a lost race. It’s that our domestic capability erodes while the argument stays fixed on the wrong thing, and that by the time we reach for it again, it’s become too slow and expensive to rebuild.