EvolutionaryScale raises $142M and launches AI model for protein discovery
EvolutionaryScale has raised $142M in a seed round led by former GitHub CEO Nat Friedman, AI entrepreneur Daniel Gross and venture capital firm Lux Capital. They were joined by Amazon, NVIDIA’s venture arm NVentures, and angel investors.
EvolutionaryScale was founded in July 2023 by former Meta employees, who are pioneers in applying AI to biology research. During their time at Meta’s Fundamental AI Research (FAIR) unit, the founders developed the first large language model for protein research, ESM1, which was able to predict the structure of hundreds of millions of proteins from metagenomics data.
The next-generation AI model launched by EvolutionaryScale, called ESM3, enables users to study and create proteins. It is the first generative AI model for biology that can analyze the sequence, structure and function of proteins simultaneously, enabling scientists to program biology and create new proteins across a wide range of applications including drug discovery, materials science and carbon capture.
ESM3 has the largest computational power of any model in biology up to date: one trillion teraflops. The AI model was trained on a dataset of 2.78 billion proteins found across planet Earth.
“ESM3 takes a step toward a future of biology where AI is a tool to engineer from first principles, the way we engineer structures, machines, and microchips, and write computer programs,” said EvolutionaryScale co-founder and chief scientist, Alexander Rives. “We’ve been working on this for a long time, and we’re excited to share it with the scientific community and see what they do with it.”
EvolutionaryScale has released a preprint paper where ESM3 was used to discover a new variant of green fluorescent protein (GFP). To date, all GFPs used in biomedical research were found in nature, making this new variant the first to be designed by humans.
“In our internal testing we’ve been impressed by the ability of ESM3 to creatively respond to a variety of complex prompts,” said Tom Sercu, co-founder and VP of engineering at EvolutionaryScale. “It was able to solve an extremely hard protein design problem to create a novel Green Fluorescent Protein. We expect ESM3 will help scientists accelerate their work and open up new possibilities — we’re looking forward to seeing how it will contribute to future research in the life sciences.”
An open version of ESM3 has been made available for non-commercial use. EvolutionaryScale is also collaborating with Amazon Web Services (AWS) and NVIDIA to accelerate applications from drug discovery to synthetic biology.
The partnership with AWS will allow EvolutionaryScale to make the full AI model accessible to hundreds of thousands of researchers around the world, as well as to nine out of the 10 top global pharma companies, which already use AWS generative AI and health services.
The computing power required to train generative AI models is growing exponentially. As part of their ongoing collaboration, NVIDIA will be providing support to EvolutionaryScale with the hardware and software necessary to handle the large amounts of data and computing power that ESM3 requires.
