For Daniil Boiko and Andrei Tyrin, an idea for Onepot artificial intelligence stemmed from the same frustration.
“The best ideas in drug discovery have often been blocked not by biology, but by synthesis,” Boiko told TechCrunch. Synthesis is the creation of latest molecules using chemical reactions. It’s like a recipe or Lego pieces, where small pieces, ingredients, molecules come together to create a larger picture of a puzzle, a dish of food, a larger molecule.
As you would possibly expect, it’s quite difficult to create these small molecules that can then build larger ones.
For Boiko, a graduate student studying machine learning in chemistry at Carnegie Mellon (he earned his bachelor’s and master’s degrees in organic chemistry from a university in Russia), this meant realizing that drug hunters – scientists who oversee drug discovery and development – were passing over promising ideas simply because the chemical molecules to make drugs seemed too difficult to produce.
“The compounds never even had a chance to be tested,” Boiko told TechCrunch.
For Tyrin (who earned a bachelor’s degree in computer science from MIT), his time working on the computational processes of drug discovery made him realize just how backward the world of drug discovery is. “Models can generate ideas in a few hours, but it can take months to catch up in the lab,” he told TechCrunch.
“We both saw that the world was pouring money into molecular design and almost ignoring the more difficult problem of actually making the molecules,” Boiko said. However, he continued, there is also a geopolitical aspect – global supply chains are becoming vulnerable and the United States is once again entering a trade war and innovation competition with China.
Techcrunch event
San Francisco
|
October 13-15, 2026
“It was clear,” Boyko said. “In the United States, small molecule synthesis had to be rebuilt from scratch.”
Boiko and Tyrin joined forces to create Onepot, the company that houses the POT-1 small molecule synthesis laboratory. They also built Phil, an AI organic chemist, to help perform experimental analyzes to improve the strategy of synthesizing compounds for their early business partners. These partners are biotechnology and pharmaceutical firms that are currently testing their technology.
The company emerged from stealth on Wednesday with $13 million in funding, including pre-seed money and a seed round led by Fifty Years.
“Currently, pharmaceutical and biotechnology companies either create entire teams of chemists in-house or contract with foreign research organizations,” Tyrin said of the molecular synthesis process. Chemists can spend months of research to create even one compound at a cost of 1000’s of dollars.
It takes a lot of trial and error – testing different compounds, collecting data on biological activity, how the drug moves through the body, toxicology reports, and determining what to experiment with next. “The main limiting factor is not the testing of these compounds, but the production of them in the first place,” Tyrin continued. “We want to get it down to days.”
Tyrin said the product is quite easy. Onepot has a catalog of molecules it will possibly produce. Customers select the compounds they need, and then Onepot technology will synthesize the molecules and send them to the customer so they’ll use them in their very own experiments. (They ship physical products as dry compounds or solutions in plates or vials.)
In the back of the product, Boiko and Tyrin have fun analyzing chemical synthesis problems to work out which mixtures of molecules work together. They have built a lab where they permit LLM agents to access so-called molecular recipes for training, so agents may also learn what works and what doesn’t in building relationships.
“When we do experiments in the lab, we record every detail of the process,” Tyrin said—which means tracking the temperature and, essentially, the ingredients added to the mixture to produce the compounds. “No information is lost, which makes the experiments repeatable even if someone decides to perform them 10 years from now.”
This also signifies that their agents generate hypotheses based on real-world experiments quite than literature data, often mined from the Internet.
Boiko called the fundraising process “hectic” and said he met his lead investor through introductions. “What was supposed to be a short meeting turned into an hours-long whiteboard session on the industrialization of synthesis,” Boiko said. Other participants in the round included Khosla Ventures, Speedinvest, OpenAI co-founder Wojciech Zaremba and Google chief scientist Jeff Dean.
The fresh capital will probably be used to build a second lab in San Francisco, allowing the team to take on more clients. It may even expand the team and its complex detection engine. When it comes to service, Boiko and Tyrin see WuXi AppTec and Enamine as competitors.
Overall, Boiko and Tyrin hope that drug discovery will probably be at least twice as fast and that perceptions will change about what is possible by harnessing “weird” chemistry that scientists once thought was inaccessible.
“Not only are you accelerating drug discovery, but you’re expanding the design space for what drugs and materials can be,” Boiko said. “A cure that we haven’t discovered yet may be out there, waiting for us to find it.”