(Yicai) Sept. 21 – Shares of Insilico Medicine climbed after the biotech firm — which leverages AI for new drug discovery and hosts its core R&D capabilities in Shanghai — said it had inked an agreement worth tens of millions of US dollars with a cutting-edge AI foundation model lab. The two parties will jointly develop, validate and commercialize a suite of advanced models tailored for biological research.

Insilico [HKG: 3696] closed 4.4 percent higher at HKD63.50 (USD8.09) a share in Hong Kong today, after earlier surging by as much as 7.2 percent.

Insilico and its partner will combine the latter’s proprietary generative foundation model architecture with the former’s MAI Gym training framework and scientific reasoning data to build collaborative and innovative foundation models for generative biology, the Boston-based company announced today.

The pair will leverage their respective global biomedical partner networks to commercialize the AI product, Insilico noted, adding that they will use a revenue-sharing mechanism for the project, without providing exact details.

Insilico will supply proprietary datasets covering multiple specialized life science tasks for the project and conduct closed-loop wet-lab validation of outputs generated during foundation model training, it said.

Insilico focuses on integrating AI and automation technologies to accelerate drug discovery and foster innovation across life sciences. Its in-house Pharma.AI platform and advanced automated biology labs deliver innovative therapeutic candidates for fibrosis, oncology, immunology, pain, obesity, and metabolic disorders, and other areas.

Since launching the MMAI Gym model training and evaluation ecosystem in January, Insilico has announced successive AI model co-development partnerships with Liquid AI, a Massachusetts-based developer of liquid foundation models, Human Longevity, a California-based firm that focuses on precision health and longevity research, and Tencent Healthcare, a unit of Chinese internet giant Tencent Holdings.

Editors: Tang Shihua, Martin Kadiev