Biotech Industry · global
GenScript Connects with Lilly TuneLab: AI-Selected Molecules Go to the Lab for Validation
GenScript will offer discounted laboratory services to TuneLab participants, connecting model predictions with protein production and testing. The collaboration has raised market expectations, but public results have yet to support whether it can improve drug candidate quality.
AI can help researchers select molecules worth trying, but experiments are still needed to answer whether they can be produced and whether they have the expected properties. On September 16, 2026, GenScript Biotech announced a collaboration with Lilly’s AI drug discovery platform, Lilly TuneLab, to offer participating companies laboratory services at discounted prices, handling the work of moving from computational predictions to biological validation.
The collaboration has also given the market a new basis for assessing GenScript. Reports from Futu Niuniu and Sina Finance on September 21 both noted that GenScript rose more than 7% during trading that day. At the time of Sina’s report, the stock was up 5.34% at HK$41.44, with trading turnover of HK$977 million. The share price reflects investor expectations for demand for AI research and development services; the collaboration’s scientific and commercial outcomes require a separate set of evidence to assess.
According to GenScript’s announcement, TuneLab gives participating companies access to AI and machine learning models trained on years of Lilly research data. Applications include predicting whether antibodies are suitable for further development, as well as the absorption, distribution, metabolism, excretion, and toxicity of small molecules. These predictions help research teams prioritize early experiments; the models’ assessments still need to be checked against actual measurements.
GenScript will handle the practical work: producing proteins from prioritized sequences, then purifying and characterizing them through standardized, fully documented procedures. Sina also included gene synthesis among the relevant services when citing a JPMorgan research report. For biotech companies using the platform, obtaining comparable experimental results is an important step in deciding which candidate molecules warrant further investment.
The arrangement also includes data feedback. GenScript said TuneLab is hosted by a third party and uses federated learning, allowing participating companies to contribute experimental results to improve the models. The research and development logic is to let predictions guide testing, then let testing refine predictions. However, the announcement provided no validation datasets, model accuracy figures, or quantified results showing improved efficiency in drug candidate evaluation for this collaboration.
The headline of Futu’s report linked the market reaction to large model vendors entering drug discovery and analysts’ assessment of a shift in GenScript’s business model. The directly verifiable development is that GenScript is integrating its existing laboratory services into TuneLab’s research and development workflow. The public announcement did not disclose the contract value or minimum purchase volumes, nor did it report any drug candidates or human trial results arising from the collaboration. Whether laboratory demand can generate recurring revenue, and whether data feedback can improve the quality of drug candidate selection, will determine how far this collaboration can go.