HONG KONG – The world’s largest pharmaceutical companies are quietly rewiring their drug-discovery pipelines around artificial intelligence platforms built in China, even as Washington moves to restrict exactly that kind of technology access.
Chinese biotech firms struck licensing deals worth $75 billion in the first five months of 2026 alone, according to reporting by South China Morning Post. Total deal value, including both in-licensing and out-licensing arrangements, reached $93 billion across 169 agreements, an 87 percent increase in deal volume from the same period a year earlier. In 2020, Chinese companies had not concluded a single out-licensing agreement with a foreign pharmaceutical partner.
The pace of deal-making accelerated even as the administration moved to tighten restrictions on technology transfers involving Chinese AI platforms. The contradiction sits at the center of one of the most consequential shifts in global pharmaceutical research in a generation: Western drug developers need what Chinese AI labs have built, and the market is moving faster than the regulators.
The deal announced last week between Metis TechBio, a Hong Kong-listed artificial intelligence drug-design firm, and Boulevard Bio, a New York biotech backed by Deerfield Management, illustrates how the pipeline has changed. Metis will receive $20 million upfront and is eligible for up to $1.6 billion in milestone payments plus royalties for rights to MTS-128, a cancer drug candidate generated by its AI platform. Metis listed on the Hong Kong Stock Exchange less than two months ago, in May 2026.
Before 2020, Chinese biotech companies recorded zero out-licensing deals with foreign counterparts. The acceleration since then has been steep enough that analysts tracking the sector now describe it in terms usually reserved for semiconductor markets rather than pharmaceutical ones. Linda Shu, HSBC’s head of China healthcare research, has pointed to the AI-driven efficiency gains in drug candidate generation as the structural factor behind the surge. Where a traditional drug-discovery program might spend years identifying a viable molecule, AI-assisted design compresses that timeline to months.
The catch is that the platforms generating those molecules are built on data and compute infrastructure that Washington is increasingly treating as a national security concern. The same AI capability that makes Metis TechBio’s MTS-128 attractive to Boulevard Bio also makes the underlying technology a target for investment-screening reviews that the administration has been expanding. That scrutiny extends beyond pharmaceuticals. A report released this week by the European Union Chamber of Commerce found European industries more structurally dependent on Chinese technology inputs than previously disclosed, a finding that has accelerated policy conversations in Brussels and Berlin about supply-chain resilience.
The tension has not yet disrupted deal flow, partly because agreements being signed now were negotiated over months or years before any new restrictions took effect, and partly because the compounds themselves, once designed, exist independently of the platforms that created them. A Chinese AI firm can license a drug candidate to a US biotech without handing over the source code or training data behind its platform. Whether that distinction will survive more aggressive regulatory scrutiny is the question the industry has not yet had to answer. The technology dependence that makes Chinese AI platforms attractive to Western pharma partners is not limited to drug design. Apple’s accommodation with Chinese chip suppliers for its next-generation devices signals how deeply US tech product cycles have become entangled with Chinese manufacturing and design capability.
What the numbers do not capture is the direction of technical dependence they reflect. European and American pharmaceutical companies are not licensing Chinese AI drug candidates because Chinese firms are cheaper; they are licensing them because the platforms are, in several measurable respects, producing better early-stage candidates faster. Metis TechBio’s Hong Kong debut in May drew institutional investor interest precisely because its pipeline was built on AI infrastructure that European companies have been slow to match. The Deerfield-backed deal with Boulevard Bio is the commercial proof of concept.
China’s biotech industry will almost certainly conclude more licensing deals in the remaining seven months of 2026 than it did in the entirety of 2025. The $75 billion figure covers only five months. If the pace holds, total out-licensing value for the year would approach $180 billion, making China the dominant source of early-stage drug candidates for the global pharmaceutical industry. Whether MTS-128 will eventually reach patients, and under what regulatory conditions, remains unknown. What is clear is that the question of who controls the AI platforms that find the next generation of cancer drugs has become inseparable from the broader contest over who controls the technology that matters most in the 21st century.
