Key Takeaways
- Adaptyv is developing automated labs that allow AI systems to design experiments and directly test protein sequences.
- The company recently raised new funding to expand its lab capacity and operations in Lausanne and London.
- Adaptyv aims to accelerate biophysical testing and broaden its experimental capabilities in drug discovery.
Automated Labs for AI-Driven Protein Design
Adaptyv, a Lausanne-based biotech company, is innovating the way proteins are designed and tested by creating automated labs that can be accessed directly by AI systems. This approach significantly bridges the gap between AI-generated designs and practical experimentation, streamlining the drug discovery process.
The recent funding round was led by Highland Europe, joined by existing investors including Ace Ventures, ByFounders, and Y Combinator. Previously, Adaptyv raised an $8 million Seed round. This new financial boost aims to expand lab capacity and operations, targeting a tripling of their capabilities by the end of 2026.
Adaptyv’s platform operates by processing digital protein sequences through various stages including DNA synthesis, protein expression, and experimental testing. This results in real-world data on the performance of the designed molecules. Users can submit designs via Adaptyv’s web platform or API, enabling AI agents to autonomously request and manage experiments without human intervention. This innovative model, termed “agentic biology,” establishes a continuous feedback loop where AI refines its designs based on experimental outcomes.
A practical application of this technology was demonstrated in collaboration with Anthropic, where the AI model Claude designed protein binders for 16 targets. Out of 1,320 designs sent to Adaptyv’s automated lab, 354 demonstrated binding success, achieving an average hit rate of 26.8%. Notably, the designs generated by Claude outperformed winning entries in five out of six comparable competitions.
Adaptyv’s efficiency has surged, reportedly increasing laboratory throughput by more than fivefold within the past year, catering to over 100 clients, which include AI-biotech firms like Chai Discovery and Boltz, as well as major pharmaceutical players like Roche and Novo Nordisk.
Looking ahead, Adaptyv plans to transition beyond protein-binding experiments to include developability studies and a wider range of biophysical tests. The company is also set to explore more complex modalities, such as peptides, antibody-drug conjugates, degraders, and eventually cell-based assays.
In an industry where much investment has prioritized advanced models and computing capabilities, Adaptyv’s strategy counters that trend by emphasizing the need for robust experimental infrastructure. Their vision includes what they describe as a “biological gigafactory,” designed to provide ample standardized experimental capacity for AI systems to continuously validate their predictions against biological realities.
The content above is a summary. For more details, see the source article.