Onepot AI raises $13M to help make chemical drug creation easier

Daniil Boiko and Andrei Tyrin, co-founders of Onepot AI, identified a significant barrier in the drug discovery process — the synthesis of chemical compounds. Boiko, a Ph.D. candidate specializing in machine learning in chemistry at Carnegie Mellon, observed that drug discovery often stalls not due to biological challenges but because of the complexities involved in creating new molecules. He explained that these molecules are the building blocks for larger compounds, akin to assembling a puzzle or preparing a dish from various ingredients. The duo realized that many promising drug candidates were overlooked simply because the synthesis of the required molecules appeared too daunting. Tyrin, equipped with a background in computer science from MIT, noted that while computational models could generate potential drug ideas rapidly, the laboratory processes lagged significantly behind, often taking months to catch up. Recognizing the necessity for innovation in small-molecule synthesis, especially in light of geopolitical tensions affecting global supply chains, they established Onepot. The company is now home to a specialized lab, POT-1, and has developed an AI-driven organic chemist named Phil, designed to streamline the experimental processes of compound synthesis for their early-stage commercial partners in the biotech and pharmaceutical sectors. Recently, Onepot emerged from stealth mode, announcing a successful funding round of $13 million, including pre-seed and seed investments led by Fifty Years. Tyrin elaborated on the traditional approach in the industry, where companies either maintain extensive in-house chemistry teams or outsource to overseas contract research organizations. This conventional method can lead to lengthy and costly research cycles, often requiring significant trial and error. Onepot aims to drastically reduce synthesis times from months to just days. Their straightforward service allows clients to select from a catalog of molecules that Onepot can produce. The company then synthesizes these compounds and ships them directly to clients for their experimental needs, whether as dry compounds or in solution. The innovative backend of Onepot’s service involves meticulous data collection during the synthesis process, ensuring that every detail, from temperature to ingredient combinations, is documented. This approach not only enhances reproducibility but also allows their AI agents to learn from real-world experiments, generating hypotheses based on empirical data rather than just theoretical literature. Amidst a whirlwind fundraising experience, Boiko described how their initial pitch evolved into an in-depth discussion about industrializing synthesis. Their investor network includes notable names such as Khosla Ventures and OpenAI co-founder Wojciech Zaremba. The new funding will facilitate the establishment of a second lab in San Francisco, enabling Onepot to scale its operations and enhance its compound discovery engine. In the competitive landscape, Boiko and Tyrin view companies like WuXi AppTec and Enamine as their main rivals. Their ultimate goal is to accelerate drug discovery processes and reshape perceptions of what is achievable in the field of chemistry. Boiko emphasized that their work not only speeds up drug discovery but also broadens the possibilities for future drugs and materials, hinting at groundbreaking discoveries still waiting to be unearthed.