Lila shifting to milestone economics

This shifts Lila from selling lab time to owning a piece of the upside when its science becomes a product. In practice, that means Catalyst can bring in service style revenue from outside R&D teams, while Creation can take the richer economics usually captured by a biotech asset owner, such as payments when a program hits technical or clinical milestones, fees for using patented discoveries, or equity in a spinout built around a new molecule, material, or platform.1 2 3 4

• This is the same basic value capture ladder used in biotech partnerships. Isomorphic Labs pairs AI discovery work with upfront payments, milestone payments, and potential royalties, and also runs wholly owned internal programs, showing how an AI science company can mix fee revenue with asset level upside.3 4
• Lila is better positioned for that model than a pure software layer because it controls both the model and the physical lab stack. That lets it generate not just predictions, but validated experimental results that can be packaged into licensable IP, partner programs, or new venture formation.1 2
• Comparable AI for science companies are converging on the same idea. CuspAI and Discovery Loop are both framed around expanding beyond software or services into licensing, milestones, royalties, or joint programs, which suggests the market is rewarding companies that can turn discovery engines into ownable assets.5

The next step is a bifurcation of business models across AI for science. Companies that only sell tools will be valued on usage and gross margin, while companies that repeatedly turn experiments into proprietary assets will start to look more like biotech or industrial IP platforms, with fewer deals but much larger outcomes per success.3 4 5