Lila Sciences leans computational biology

This points to a company whose core edge sits in turning biological data into predictions, not in building a general purpose robot lab from scratch. Recursion’s biggest assets are BioHive-2, its imaging pipelines, and the software stack that learns from huge cell datasets, while Lila is trying to make autonomous experimentation itself the product, across biology, chemistry, and materials science.

• Recursion has invested heavily in compute and data. It built BioHive-2 with NVIDIA and uses robotized imaging to generate large multimodal datasets for small molecule discovery. That is a computational biology loop, where robots mainly serve data collection for model training and screening.
  2 recursion 3 recursion
• Lila is organized around autonomous lab operations more directly. Its AI Science Factory is designed so models generate experiments, instruments run them, data comes back automatically, and the next experiment is redesigned in a closed loop. That makes lab orchestration a primary capability, not just support infrastructure.
  1 sacra 4 lila
• Insilico sits between the two. It combines target discovery and generative chemistry software with automated labs, but its public materials still frame robotics as an extension of its AI drug discovery platform. That makes Recursion and Insilico closer to software first drug discovery companies, while Lila is pushing toward a fuller autonomous science stack.
  5 insilico 6 insilico 7 insilico

The market is moving from AI assisted discovery toward fully closed loop science systems. Companies with strong biology models will keep adding automation, but the next wedge is owning the full cycle from hypothesis to experiment to new training data. That favors platforms like Lila that treat autonomous lab execution as core infrastructure, not a downstream tool.