EvolutionaryScale develops biological artificial intelligence models, specifically ESM3, to generate and design proteins by analyzing sequence, structure, and function data. This technology enables scientists to create novel proteins, such as esmGFP, significantly faster than natural evolutionary processes, enhancing therapeutic design capabilities in life sciences.
Funding
$182M raised to dateRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
NFFounders
Product
Problem
The process of designing and generating novel proteins with specific functions is traditionally slow and resource-intensive, often taking years using directed evolution or rational design approaches. Scientists need faster methods to create proteins for various applications, including therapeutics, industrial enzymes, and biomaterials.
Solution
EvolutionaryScale provides ESM3, a family of large-scale AI models that enable scientists to generate and design proteins with unprecedented speed. By training on a massive dataset of protein sequences, structures, and functions, ESM3 can simultaneously reason over these fundamental properties to explore a vast space of possibilities. Users can input a mix of sequence, structure, or function data to prompt ESM3 to create novel proteins, such as esmGFP, a green fluorescent protein that would have taken nature 500 million years to evolve.
Target Audience
The primary users are scientists and researchers in life sciences, biotechnology, and pharmaceuticals who need to design and generate novel proteins for various applications, including therapeutics, industrial enzymes, and biomaterials.
Features
- Generative AI models trained on billions of natural proteins to understand the language of biology
- Chain-of-thought prompting allows users to specify desired sequence, structure, and function properties
- Ability to design proteins with vast evolutionary departure from natural proteins
- Available in three sizes (small, medium, and large) to suit different computational needs
- Accessible through an API and partner platforms like Forge, AWS (Sagemaker, Omics platform, Bedrock), and NVIDIA BioNemo
- ESM3-open, a small but powerful model, is available on GitHub under a non-commercial license
- Adherence to a responsible AI development framework ensures safe and ethical use