🔬 The Lab of the Future Should Feel Like a Data Center — Andy Beam & Rafa Gómez-Bombarelli, Lila Sciences

Lila Sciences is building an automated laboratory designed to function like an AI data center. By integrating AI-guided robotics with physical science experiments, the company aims to create a scientific superintelligence capable of solving complex problems in biology, chemistry, and materials science. The facility operates 24/7, using a magnetically levitating transport system to move samples between instruments, which are treated as nodes in a computational network.

The core of this approach is the generation of massive amounts of experimentally validated data. Lila has produced over 10 trillion scientific reasoning tokens, which serve as the foundation for training general-purpose models. Rather than focusing on high-throughput screening, the company prioritizes rapid, iterative cycles of experimentation. By treating the scientific method as a data generation mechanism and nature as the ultimate verifier, they aim to move beyond domain-specific models to create a system that can reason across different scientific disciplines.

This model has already yielded practical results, such as identifying high-performing electrocatalysts and accelerating the development of in vivo CAR-T therapies. By automating the scientific process, Lila seeks to replace serendipitous discovery with systematic, scalable reasoning. This strategy aims to eliminate the reliance on luck in scientific breakthroughs, potentially transforming how new drugs and materials are discovered by treating the lab as an infinite generator of verified scientific knowledge.