Lila Sciences bets AI runs the whole scientific method: $550M raised, $1.3B valuation

A Cambridge, MA company building a 'scientific superintelligence' platform and AI Science Factories: automated robotic labs that generate hypotheses, design and run experiments, and learn from results across life sciences, chemistry, and materials.

That the scientific method — hypothesis, experiment, learn — can be automated end-to-end by AI plus robotic labs, so discoveries that took labs years now take weeks.

$235M initial round expanded and closed at $350M Series A; $550M total raised; ~$1.3B valuation (Aug 2026)

Flagship Pioneering founded Lila Sciences in 2023, with Geoffrey von Maltzahn, a former Flagship partner, as co-founder and CEO. Flagship had already spun out Moderna, and Lila's premise was that the scientific method itself — not any single domain — is what AI can automate.

Lila demonstrated early proof: AI-generated novel antibodies, carbon-capturing materials, and water-splitting catalysts that outperform the precious metals industry relies on — a third of which its Caltech-trained team initially dismissed, and which Fortune reported are now the highest-performing catalysts on record. Its first round grew from $235M to a $350M Series A closed in mid-2026 with NVentures (NVIDIA), Analog Devices, and others; total raised hit $550M at about a $1.3B valuation. It signed a 235,500 sq ft facility lease in Cambridge and began welcoming its first cohort of customers.

Scaling: $550M raised at ~$1.3B valuation, CNBC Disruptor 50 #25, first commercial customers in onboarding, and AI-designed water-splitting catalysts reported as the highest-performing on record.

The scientific method is the same loop in every field, so automating it once creates a platform that spans drugs, materials, and energy instead of a single product.

Flagship's incubation gave Lila the capital and runway to build AI models and physical labs together, which pure-software startups cannot.

Demonstrated wins in multiple domains — antibodies, carbon capture, catalysts — turned a philosophical bet into evidence investors could price.

The $1.3B valuation while still onboarding its first customers shows the market priced the platform, not the revenue.

A 2,400% search-growth month indicates the bet itself became the story people were looking up.

Automate the loop, not the product: Lila owns hypothesis, experiment, and learning across domains, so a breakthrough in one field becomes software IP that compounds into the next.

As of Aug 18, 2026, Lila had closed its Series A at $350 million for $550 million total, a valuation around $1.3 billion, a 235,500 sq ft Cambridge facility lease, and its first cohort of commercial customers in onboarding. Von Maltzahn was publicly arguing that agentic AI and the scientific method are structurally the same loop, and that the 'Claude Code moment' — when AI permanently changes how science is done — is imminent.