Eli Lilly Makes its AI Models Available to Biotech Sector via Benchling
Eli Lilly is making its AI/ML platform, TuneLab, broadly accessible through an integration with San-Francisco based Benchling, bringing models trained on over $1 billion worth of proprietary research data into the hands of more than 1,300 biotech companies. The collaboration is a response to the longstanding limitations in AI model training across life sciences, particularly lack of high-quality data and secure infrastructure for model deployment.
Full access to TuneLab within Benchling is expected to roll out later in 2026.
TuneLab, first launched in September 2025, is Lilly’s federated learning platform for early-stage drug discovery, offering access to AI models trained on proprietary preclinical, safety, and molecular datasets, including ADME, toxicology, and PK/PD profiles from hundreds of thousands of compounds.
Through this integration, users can apply Lilly's predictive models for antibodies and small molecules directly within Benchling’s research environments, and contribute back anonymized performance data via federated learning mechanisms. The approach is designed to enhance model generalizability without centralized data pooling, preserving intellectual property (IP) and compliance controls.
Benchling is a cloud-based R&D platform built for the life sciences to address the gap between rapid advances in biology and outdated digital infrastructure. Benchling’s 2026 Biotech AI Report, which includes input from leaders at Bristol Myers Squibb, Takeda, Anthropic, and OpenAI, found that fewer than 25% of emerging biotech companies currently participate in data-sharing frameworks, even though 81% of early AI adopters have already deployed models in their research pipelines. The report identifies data availability, IP protection, and model validation as major bottlenecks for scale. We regularly cover updates like this in our weekly Where Tech Meets Bio newsletter.
Lilly selected Benchling as its primary distribution platform due to its reported widespread use among life sciences R&D teams and existing integrations with other AI toolsets, including Anthropic and NVIDIA.
This integration adds momentum to broader efforts within biopharma to expand access to industrial-scale models for small and mid-size biotech firms, including the Elix–LINC federated platform built on data from 17 pharma companies, AbbVie and J&J training the OpenFold3 structural biology model via federated data sharing, and Novartis opening its AI innovation hub to external biotechs, as well as Baishenglai (BSL), an open-access deep-learning platform for virtual drug discovery that unifies seven core tasks (including generative design and graph neural networks) under one modular system, introduced in China.
Topic: Industry Movers
