Former Meta researchers have launched a new biotech startup called EvolutionaryScale, securing $40 million in funding from Lux Capital. Led by Alexander Rives, the team aims to build AI models for biology that can revolutionize drug development and biotech applications. Their previous work involved developing an AI language model for biology, which they used to predict the structures of proteins and create a database of 700 million possible 3D structures. With this new funding, EvolutionaryScale plans to significantly scale up their AI model, potentially surpassing the capabilities of DeepMind’s AlphaFold. The long-term goal is to create a general-purpose AI model that can be applied to various biological applications, from developing cancer-fighting cells to cleaning up pollution. However, they acknowledge that it may take up to ten years for their models to have a significant impact. Despite competition and skepticism, EvolutionaryScale is confident in the scalability of their AI model and aims to prove its advantage over existing solutions.
Ex-Meta Researchers Launch AI Biotech Startup
EvolutionaryScale, a new AI biotech startup, has been launched by former Meta researchers who have raised a significant amount of funding. Led by Alexander Rives, the startup has secured $40 million in funding to support its ambitious goals. The team at EvolutionaryScale consists of former Meta researchers who have created a transformers-based AI model that has been trained on protein molecules. By leveraging this model, the team has been able to build a database with 700 million possible 3D structures, which has immense potential in the field of biotech and medicine.
In June, the startup pitched for seed financing to advance its research efforts. Lux Capital led the round, and the startup is now valued at $200 million. Notable AI investors, Nat Friedman and Daniel Gross, also participated in the funding. The focus of EvolutionaryScale is to predict protein structures, which is of great importance in the field of biotech. With their innovative AI model, the startup aims to revolutionize drug development and other biotech applications.
Importance of Predicting Protein Structures
Understanding protein structures is crucial in various fields, such as medicine and biotech. Proteins are essential molecules that play a significant role in the functioning of cells and organisms. The shape of a protein is intricately connected to its function, and it can change when it interacts with other chemicals or proteins in the body. Predicting protein structures is a complex task but holds great potential for designing drugs and developing therapies. By accurately predicting the structure of proteins, scientists can better understand their functions and develop targeted treatments for diseases.
One notable advancement in protein structure prediction has been made by DeepMind’s AlphaFold system. DeepMind’s AI system has been able to predict protein structures with remarkable accuracy and has been hailed as a significant breakthrough in the field. Comparing EvolutionaryScale’s model to AlphaFold, the startup’s model shows incremental improvements in drug development efficiency. The use of AI in predicting protein structures has the potential to greatly enhance the efficiency and effectiveness of drug discovery and development processes.
Meta AI’s Focus Shift and Layoffs
Meta, the tech giant that housed the researchers behind EvolutionaryScale, underwent a focus shift that affected its AI research efforts. Rives’ team departed Meta in April, as the company began to narrow its focus on more commercially viable projects. The AI for biology sector has limited business returns in the short term, leading to layoffs and a shift in focus at Meta and other companies in the industry. While the potential of AI in biology is immense, it is still a nascent field that requires substantial investment and research to realize its full potential.
Other companies have also been raising capital for transformer-based AI research, focusing on applications in biology and healthcare. Stability AI, for example, has faced challenges in revenue generation despite raising a significant amount of funding. The field of AI in biology is still in its early stages, and companies are grappling with the complexities and uncertainties of commercializing generative AI models.
Investment and Challenges in Protein Folding AI
Predicting protein structures and advancing AI models for protein folding is a significant challenge that requires substantial investment and research. DeepMind, recognizing the importance of this field, established Isomorphic Labs, a new drug discovery arm focused on leveraging AI for protein folding. Insitro and Recursion, two other notable companies in the space, have raised over $1 billion to support their research in this area. The average time it takes for a drug to receive FDA approval is around 7-10 years, highlighting the complex and rigorous process involved in drug development.
EvolutionaryScale projects significant spending on its AI model, with a focus on scaling it to achieve a breakthrough in AI for biology. The company aims to invest heavily in computing power, with projected spending of $38 million in its first year alone. Scaling the AI model, increasing its size and feeding it more data, is key to reaching new capabilities in the field. The long-term vision of the company is to develop a general purpose AI model for biology that can be applied to various use cases beyond protein structure prediction.
Competition in the field of protein folding AI is intense, with DeepMind’s AlphaFold being the dominant player. However, other companies, such as Inceptive, are also emerging to tackle similar goals. The potential of large language models, like OpenAI’s GPT series, being applied to biology is also worth considering. The field of AI in biology remains dynamic and full of possibilities, with various companies and researchers pushing the boundaries of what is possible.
Future Plans of EvolutionaryScale
EvolutionaryScale has ambitious plans for the future beyond protein structure prediction. The startup aims to integrate other biological data, such as DNA sequences, gene expression, and epigenetic states, into its AI model. This expansion will allow for a more comprehensive understanding of biological systems and open up new possibilities for research and development. The long-term vision of EvolutionaryScale is to sell a general purpose AI model for biology, which can be applied across various domains, including medicine and biotech.
However, EvolutionaryScale may face potential competition from other companies, such as Inceptive, that are also pursuing similar goals. The field of AI in biology is rapidly evolving, and the market is becoming increasingly competitive. Despite these challenges, the potential applications of AI models in medicine and biotech are vast. From developing programmable cells for targeted cancer therapies to designing molecular machines for environmental cleanup, AI has the potential to revolutionize these fields and drive significant advancements.
As for Alexander Rives, the interim CEO of EvolutionaryScale, his future plans include joining the faculty at MIT and Harvard’s Broad Institute. Rives aims to continue his research and contribute to the field of biological design. His expertise and experience will undoubtedly be valuable in advancing the goals of EvolutionaryScale and pushing the boundaries of AI in biology.
In conclusion, EvolutionaryScale, led by Alexander Rives, is an AI biotech startup that has raised significant funding to advance its research in protein structure prediction. The startup focuses on leveraging AI models trained on protein molecules to predict protein structures. The importance of predicting protein structures lies in its potential to revolutionize drug development and other biotech applications. While the field of AI in biology faces challenges and intense competition, companies like EvolutionaryScale are pushing the boundaries and exploring the vast potential of AI in these domains. With the right investments and research, AI has the power to transform medicine and biotech, leading to groundbreaking advancements and improved healthcare outcomes.