Anthropic Acquires Mithril Medicine, One-Person Pipeline Model Drives CRO Demand

Article body

Recently, Anthropic, which is approaching its IPO, outlined a $2 trillion pharmaceutical market opportunity in its prospectus, while also officially announcing long-term partnerships with multiple pharmaceutical companies.

This move comes just one month after it announced the establishment of a life sciences research team and laboratory. The giant's continuous actions have once again heated up the imagination around AI entering the life sciences track. Just as the market hotly debates whether AIDD will become the next killer AI application, the R&D organizational paradigm of pharmaceutical companies is also quietly being reshaped in this wave.

Just this week, a new drug R&D model called "Pipeline-in-a-Person" emerged. This model is similar to OPC (One Person Company): the founder is the company's only full-time employee, and with only AI capabilities and external laboratories, they can advance a theoretically unlimited number of pipeline experiments.

The practitioner of this model is AI pharmaceutical company Mithril Medicine, whose acquirer is none other than Anthropic. On October 27 local time, the latter disclosed the deal and claimed that the former's related projects could enter human clinical trials as early as late 2028.

Mithril Medicine founder Cody Tranbarger is a former Goldman Sachs healthcare analyst. In his view, joining Anthropic allows the original projects to continue advancing and also provides more resources, infrastructure, and fully mobilizable Claude support. In fact, with Claude's support, the company's research on a treatment for a certain rare muscular dystrophy has already entered the animal testing stage.

▌The "shovel sellers" step into the spotlight

The answer is probably no. A technical report disclosed by Mithril Medicine shows that because this one-person pharmaceutical company does not build its own laboratory, CROs (Contract Research Organizations) have instead become the center of gravity of the entire new drug R&D process.

According to statistics, in this 38-page technical report, the term CRO is mentioned 58 times, while AI/Claude is mentioned only 47 times. Within just one week, 5 CROs with different functional divisions participated in 113 molecular constructs awaiting screening, covering the full chain of experiments from antibody cell-level functional validation, drug developability assessment, to mouse pharmacokinetics.

As Anthropic acknowledges, handling biological and chemical materials and complying with biosafety and environmental requirements are beyond its existing experience as a software and AI research organization. Mithril Medicine also admits that the current pace of project advancement is not constrained by AI molecular design capabilities, but is highly constrained by CROs' experimental delivery efficiency and researchers' offline decision-making speed.

In the current market context, CROs are regarded as the "shovel sellers" of the AI pharmaceutical industry. A Caitong Securities research report states that the AI pharmaceutical transmission chain has entered a positive cycle, and the model of accelerating design, then wet-lab validation, then feeding back into AI model iteration has been preliminarily established, benefiting upstream protein synthesis, preclinical CRO and other segments first.

Zheshang Securities stated that the spillover of AI pharmaceutical bottlenecks means demand will be transmitted to the physical experiment side, and upstream "shovel sellers" will benefit first. Frontier AI institutions continue to increase their investment in life sciences, driving DNA synthesis toward multi-sequence parallel upgrades, growing demand for protein and molecular enzyme reagents, and preclinical CROs will receive upward orders from the volume ramp-up of AIDD.

▌The performance realization cycle has arrived

Amid this AI pharmaceutical wave, the domestic CRO industry has gradually recovered.

As the domestic third-quarter earnings season nears its end, nearly 30 A-share CRO companies (including pan-CXO companies involved in CDMO business) have successively delivered their performance reports: according to Statistics from the STAR Market Daily, after excluding companies with smaller market capitalization (below 10 billion yuan), the remaining 13 almost all achieved year-on-year double growth in single-quarter revenue and net profit attributable to parent companies.

Among them, Innostellar topped the list with a performance growth rate of more than 100 times. The company stated that in the first three quarters of 2026, the company's newly signed order amount was 2.218 billion yuan, a year-on-year increase of 183.27%; as of the end of the reporting period, the company's order backlog was 2.625 billion yuan, an increase of 110.34% over the end of 2025. The rapid thickening of newly signed and backlog orders fully reflects the synchronous improvement of market demand prosperity and the company's order acquisition capability.

In addition, WuXi AppTec, Pharmaron, Asymchem, Porton Pharma Solutions and others all disclosed order growth expectations in their third-quarter reports:

WuXi AppTec stated that as of the end of September 2026, the company's order backlog for continuing operations was RMB 76.74 billion, a year-on-year increase of 36.4%. It expects the company's overall revenue for 2026 to be raised from RMB 58.5-60.5 billion to RMB 61.0-61.9 billion, and the year-on-year growth rate of revenue from continuing operations to be raised from 35%-39% to 40%-42%.

Pharmaron stated that in the first three quarters of 2026, the company's newly signed orders increased by more than 30% year-on-year. Among them, newly signed orders in the laboratory services segment increased by more than 25% year-on-year, newly signed orders in the CDMO services segment increased by more than 40% year-on-year, and newly signed orders in the clinical research services segment increased by more than 25% year-on-year.

Asymchem stated that, judging from the company's backlog of orders awaiting delivery, the delivery scale in the fourth quarter will increase significantly compared with this period, and the company expects full-year 2026 operating revenue to grow by 19%-22%.

At the same time, companies are successively aligning with the AI narrative. WuXi AppTec mentioned AIDD/AI4S for the first time in its third-quarter report, stating that it will efficiently meet the rapidly growing dry-wet closed-loop R&D needs of global AIDD/AI4S customers; Pharmaron stated during an institutional research survey on October 28 that the company's bioscience segment grew faster than laboratory chemistry, and the acceleration in the third quarter was even more pronounced, one of the driving factors being that it seized business opportunities brought by new AI-enabled companies.

A research report released by CITIC Securities states that AIDD companies are generally in an early stage, and the industry is expected to enter an accelerated growth period. AI biotech mainly includes three business models: first, SaaS-type companies that only provide software services; second, service-type CRO companies that combine AI platforms to screen or evaluate new drugs for customers; and finally, AI innovative drug companies that independently design new drugs through AI platforms and advance pipelines.

The institution stated that most companies are currently in an early stage of development, adopting a combination of several of these business models. The overall AI pharmaceutical market is still relatively small, but with the participation of more and more capital, technological iteration, and the gradual listing and validation of AI pharmaceutical pipelines, the industry is expected to gradually usher in a harvest period.

(Source: STAR Market Daily) Body middle Embedded advertisement