For Daniil Boiko and Andrei Tyrin, the idea for Onepot AI came from the same frustration.
“The best ideas in drug discovery were often blocked by synthesis rather than biology,” Boyko told TechCrunch. Synthesis is the use of chemical reactions to create new molecules. It's like a recipe or a Lego piece, where smaller pieces, ingredients, and molecules come together to form broader puzzle pictures, food plates, and larger molecules.
As you can imagine, creating small molecules to build larger molecules is extremely difficult.
In Dr. Boyko's case, he was a candidate studying machine learning in chemistry at Carnegie Mellon University (he earned his bachelor's and master's degrees in organic chemistry at a Russian university). It meant that drug hunters (scientists who oversee the discovery and development of new drugs) found themselves ignoring promising ideas simply because it was too difficult to create the chemical molecules to make the drugs.
“The compound wasn't even given a chance to be tested,” Boyko told TechCrunch.
For Tyrin (who earned a bachelor's degree in computer science from MIT), working on computational pipelines for drug discovery made him realize how far behind the world of drug discovery was. “Models can generate ideas in hours, but it can take months for labs to catch up,” he told TechCrunch.
“We both saw the world pouring money into designing molecules and largely ignoring the more difficult problem of actually making them,” Boyko said. However, there is also a geopolitical dimension, with global supply chains becoming increasingly fragile and the United States once again entering a trade war and innovation competition with China, he continued.
tech crunch event
San Francisco | October 13-15, 2026
“It was obvious,” Boyko said. “We had to rebuild small molecule synthesis from scratch in the United States.”
Boyko and Tyrin teamed up to found Onepot, a company based in the small molecule synthesis lab POT-1. We also built Phil, an AI organic chemist, to help our early commercial partners perform experimental analyzes to improve their compound synthesis processes. These partners are biotech and pharmaceutical companies currently trialling their technology.
On Wednesday, the company came out of stealth with $13 million in funding, including pre-seed funding and a seed round led by Fifty Years.
“Right now, pharmaceutical and biotech companies are building entire teams of chemists in-house or collaborating with contract research institutions overseas,” Trinh said of the process of molecule synthesis. Human chemists can spend months of research and cost thousands of dollars to create a single compound.
It requires a lot of trial and error, studying different compounds, gathering data on biological activity, how drugs move through the body, toxicology reports, and figuring out what to experiment next. “The main limiting factor here is not testing these compounds, but making them in the first place,” Tyrin continued. “We aim to reduce this to just a few days.”
Tyrin said the product is very easy to use. Onepot has a catalog of molecules that you can create. Once a customer selects a desired compound, Onepot's technology synthesizes the molecule and ships it to the customer for use in their own experiments. (The physical product is shipped as a dry mixture of solutions in plates or vials).
The back end of the product is where Boyko and Tyrin have fun analyzing chemical synthesis problems and seeing what combinations of molecules work together. They built a lab that gave LLM agents access to these so-called molecular recipes for training, allowing agents to see what works and what doesn't in building compounds.
“When we perform experiments in the lab, we record every detail involved in the process,” Tyrin said. This means tracking the temperature and basically the ingredients added to the mixture to create the compound. “No information is lost, so if someone decides to run it 10 years from now, they can still recreate the experience.”
This also means that their agents often generate hypotheses from real-world experiments rather than literature data mined from the internet.
Boyko called the fundraising process “hectic” and said he met the lead investor through an introduction. “What was supposed to be a short meeting turned into a multi-hour whiteboard session about industrializing synthesis,” Boyko said. Other participants in the round include Khosla Ventures, Speedinvest, OpenAI co-founder Wojciech Zaremba, and Google Principal Researcher Jeff Dean.
The new funding will be used to build a second lab in San Francisco, allowing the team to reach more customers. It will also expand the team and its compound discovery engine. On the services side, Boiko and Tyrin consider WuXi AppTec and Enmine to be competitors.
Overall, Boyko and Tyrin hope to at least double the speed of drug discovery and change perceptions of what's possible after harnessing “weird” chemistry that scientists once considered off-limits.
“It not only accelerates drug discovery, but expands the design space of what drugs and materials can be,” Boyko said. “There may be drugs out there that we haven't discovered yet, waiting for us to discover them.”