The Biotech Platform That Turned Profitable — and What It Means for the Industry

Sanyou Bio made AI drug discovery profitable. Now it wants to rent out the machine that made that possible — and that is where the story gets complicated.

The Shanghai company, founded in 2015, announced Thursday it has built a profit base with sustained positive cash flow in 2025, rare enough in AI drug discovery to be worth noting carefully LinkedIn retrospective by founder Lang Guojun. It also launched Clicklinks, a portal that integrates nearly 70 standardized technology modules across six gateways covering core reagents, drug discovery, pharmacology, preclinical R&D, manufacturing, and disease focus areas, and says any pharma company, biotech, academic lab, or CDMO can plug into its infrastructure rather than building from scratch PR Newswire. No external partners have publicly committed yet. The figures are self-reported on LinkedIn. An audited filing or a disclosed partnership term sheet has not surfaced publicly.

Here is the part that should make pharma lawyers reach for their keyboards. Traditional drug discovery protects value through molecule exclusivity: you discover a compound, you patent it, you control who makes it. Sanyou is proposing a different model — rent access to a discovery infrastructure that already contains a 10-trillion-molecule antibody library, AI acceleration tools, and standardized module layers PR Newswire. The pitch is velocity. The complication is the IP gray zone that follows any molecule discovered through shared infrastructure: who owns what, and under what exclusivity terms?

That question does not have an obvious answer inside the current pharma IP framework. The patent system was designed for molecules, not for discovery workflows that span multiple platforms and legal entities. If distributed discovery infrastructure scales, the entire licensing and exclusivity architecture that underpins biotech company valuations, platform-promise pitches, and pharma partnership deals needs rethinking. Whether that rethinking happens because of Sanyou, or whether Sanyou itself runs into the wall before the model proves out, is the open question.

The mechanism behind Sanyou's claim to profitability is a scale bet that has taken a decade. The company built a 10-trillion-molecule antibody library called AI-STAL — at 10^13, it says this is 100 to 1,000 times larger than most domestic libraries, which typically sit in the 10^10 to 10^11 range LinkedIn retrospective. The library is the raw material for three platforms: AI-STAL for library and target search, SAI-DA for AI-driven drug acceleration, and now Clicklinks as the infrastructure layer on top.

Lang founded the company after earning his PhD from Zhejiang University, betting that antibody library scale would be the constraint nobody else was willing to solve. In 2015 the company started with a 10^10 library in a 100 square meter lab. Today it operates more than 20,000 square meters of R&D and GMP facilities in Shanghai Dealroom.co. The growth in between was a decade of building infrastructure that the current platform stack sits on top of.

The SAI-DA accelerator, which the company launched publicly in early 2026, compresses lead discovery from three to six months down to 14 days, and takes the path from initial compound to preclinical candidate from one to three years down to two to three months LinkedIn retrospective. Clicklinks adds the portal layer on top — a portal integrating six gateways covering core reagents, drug discovery, pharmacology, preclinical R&D, manufacturing, and disease focus areas, supported by nearly 70 standardized technology modules PR Newswire.

The collaboration numbers are more concrete than the financial claims. Sanyou has completed more than 50 collaboration projects, more than 10 of which produced INDs that entered clinical development PR Newswire. It has filed 170 or more patent applications Dealroom.co. It says it has worked with 2,000 or more pharmaceutical companies globally across 1,200 or more drug discovery projects Dealroom.co. Those figures appear in press releases and the Dealroom database, and they are consistent with the kind of volume a platform model generates — more shots on goals from a larger library.

The China-specific barriers to export are real. Domestic R&D labor costs are structurally lower. The regulatory environment has specific pathways — NMPA approval, China-specific trial requirements — that Western pharma cannot simply replicate. Patient population access for domestic trials creates data advantages that require being inside the system. Sanyou's model works in Shanghai; whether it works when licensing to a German biotech or a Boston pharma company is a different question, one that depends on whether the library advantage survives contact with Western IP frameworks, contract research norms, and the skepticism of buyers who have heard platform pitches before.

The IP question is not academic. If distributed discovery infrastructure becomes a credible alternative to in-house build, the licensing deals, partnership terms, and VC models that assume molecule-level exclusivity need renegotiation. That is the earthquake. Whether Sanyou is the company that forces it, or whether it runs into the same IP wall before the model proves replicable, is the story between the lines.

Sanyou's 10^13 antibody library claim is large but not outside the range of what major global platforms have reported at scale. Whether the number holds matters because it is the source of the velocity advantage: a larger library means more starting points, fewer wet-lab rounds, and lower cost per candidate. If the library is smaller than claimed, the economics narrow accordingly.

What is not in dispute is that the company is still building. Clicklinks is new. SAI-DA is months old. The cash flow claim covers a single year. None of this is a verdict. It is a direction.