EvolutionaryScale releases AI models called ESM3 to help engineer novel proteins and raised a $142M seed led by Nat Friedman, Daniel Gross, and Lux Capital
Context & Ripple Effects
EvolutionaryScale emerged from Meta's discontinued protein-folding team with an earlier seed round exceeding $40 million, positioning the company around large language models for biology. ESM3 and the new $142 million financing turn that founding thesis into a more concrete product-and-capital milestone.
The move lands in an AI protein-engineering field that already included Cradle's $24 million Series A for faster protein design, making model development, rather than funding alone, a central point of differentiation.
First-order effects
- EvolutionaryScale gains $142 million to support development and commercialization around ESM3, while the release gives researchers and prospective industry users a named model platform for novel-protein engineering.
- Nat Friedman, Daniel Gross, and Lux Capital become the prominent financial backers of a biology-focused AI company at a substantially larger scale than EvolutionaryScale's earlier seed financing.
Second-order effects
- Other AI protein-design companies face a clearer competitive benchmark: they must differentiate through model performance, biological-data access, workflow integration, or customer adoption rather than a general AI-for-biology pitch.
- The size and visibility of the round can raise the bar for early-stage computational-biology teams seeking capital, while increasing investor attention on protein-engineering platforms with credible technical lineages.
Third-order effects
- If similarly funded model releases continue, protein engineering may consolidate around a smaller set of well-capitalized foundation-model builders, with specialized companies competing on applications and experimental feedback loops.
- The development is one data point in the financialization of AI-enabled biological R&D: capital is increasingly backing model platforms before their downstream commercial use cases are fully settled.
The trend: AI protein engineering is moving from venture-backed research teams toward capital-intensive model platforms competing to become core infrastructure for biological design.