← All companies

Science / Company profile

EvolutionaryScale

What EvolutionaryScale does

EvolutionaryScale (website: evolutionaryscale.ai) is an AI-for-science company focused on biological foundation models and generative protein design. Its flagship model family is ESM3, which EvolutionaryScale describes as a generative “language model” for proteins that can reason over multiple biological modalities—protein sequence, structure, and function—and generate novel proteins. The company also introduced ESM Cambrian as a parallel family aimed at improved representation learning efficiency for protein sequence modeling. EvolutionaryScale makes ESM3 available through an API on its Forge platform (closed beta as described on its site) as well as via partner deployment routes including AWS, and it also provides an open model variant under a non-commercial license (ESM3-open).

News

Company record

Source map · 2 recurring channels · 20 references

Still resolving: Newsroom · Official X

Funding

Latest disclosed valuation$200M
Tracked capital$182M3 sourced rounds
DateRoundRaisedValuationLead / investorsEvidence
Nov 06, 2025Strategic transaction / acquisition (terms undisclosed)
Aug 25, 2023Seed$40M$200M

What it builds

ESM3

ESM3 is EvolutionaryScale’s flagship generative protein foundation model family that (per the company) can simultaneously reason over protein sequence, structure, and function and generate novel proteins. EvolutionaryScale describes ESM3 availability via its Forge API in closed beta, and it also describes an open model variant (ESM3-open) with weights and source code released under a non-commercial license.

source ↗
Forge

Forge is EvolutionaryScale’s developer platform/API used to access ESM3 models (described on the company website as being available through Forge in closed beta, with an API access flow).

source ↗
ESM Cambrian

ESM Cambrian is a parallel protein-model family introduced by EvolutionaryScale, described as a new frontier for protein sequence modeling with improved performance and efficiency for representation learning, alongside ESM3.

source ↗

Key metrics