For Daniil Boiko\u00a0and\u00a0Andrei Tyrin,\u00a0the\u00a0idea for\u00a0Onepot\u00a0AI came from\u00a0the\u00a0same frustration.\u00a0\u00a0<\/p>\n
\u201cThe\u00a0best ideas in drug discovery were often blocked not by biology, but by synthesis,\u201d\u00a0Boiko\u00a0told TechCrunch.\u00a0Synthesis is\u00a0the\u00a0creation of new molecules by using chemical reactions.\u00a0It\u2019s\u00a0like a recipe or Lego pieces, where small pieces, ingredients, molecules, come together to form a wider puzzle picture, a food dish, a\u00a0larger molecule.\u00a0\u00a0<\/p>\n
As one might expect,\u00a0it\u2019s\u00a0quite hard to create those small molecules that go on to build bigger ones.\u00a0\u00a0<\/p>\n
For Boiko,\u00a0a\u00a0Ph.D.\u00a0candidate studying machine learning in chemistry at\u00a0Carnegie Mellon\u00a0(he\u00a0received\u00a0his bachelor\u2019s and master\u2019s in organic chemistry from\u00a0a\u00a0university in\u00a0Russia), that meant\u00a0realizing that drug hunters \u2014\u00a0the\u00a0scientists who oversee drug discovery and development \u2014 were skipping promising ideas just because\u00a0the\u00a0chemical molecules to create\u00a0the\u00a0drugs seemed too hard to make.\u00a0\u00a0<\/p>\n
\u201cThe\u00a0compounds never even got a chance to be tested,\u201d Boiko\u00a0told TechCrunch.\u00a0\u00a0<\/p>\n
For Tyrin (who received his\u00a0bachelor\u2019s\u00a0in computer science at MIT), his time working\u00a0on drug discovery computational pipelines made him realize how behind\u00a0the\u00a0world of drug\u00a0discovery\u00a0was. \u201cThe\u00a0models could generate ideas in\u00a0hours,\u00a0but it could take months for\u00a0the\u00a0lab to catch up,\u201d he told\u00a0TechCrunch.\u00a0\u00a0<\/p>\n
\u201cWe both saw that\u00a0the\u00a0world was throwing money into molecular design and almost ignoring\u00a0the\u00a0harder problem of actually making\u00a0the\u00a0molecules,\u201d\u00a0Boiko\u00a0said.\u00a0But\u00a0there was\u00a0a geopolitical angle too,\u00a0he continued, global supply chains are becoming\u00a0vulnerable,\u00a0and\u00a0the\u00a0U.S. is entering a trade war and innovative competition, once again, with China.\u00a0\u00a0<\/p>\n
Techcrunch event<\/p>\n
\n\t\t\t\t\t\t\t\t\tSan Francisco
\n\t\t\t\t\t\t\t\t\t\t\t\t\t|
\n\t\t\t\t\t\t\t\t\t\t\t\t\tOctober 13-15, 2026\n\t\t\t\t\t\t\t<\/p>\n
\u201cIt was clear,\u201d\u00a0Boiko\u00a0said.\u00a0\u201cSmall\u00a0molecule synthesis needed to be rebuilt from\u00a0the\u00a0ground up in\u00a0the\u00a0United States.\u201d\u00a0<\/p>\n
Boiko and Tyrin came together\u00a0to create\u00a0Onepot, a company that is home to\u00a0the\u00a0small-molecule synthesis lab POT-1.\u00a0They also built Phil, an AI organic chemist, to help run experimental analysis to increase\u00a0the\u00a0process\u00a0of compound\u00a0synthesis\u00a0for\u00a0their early commercial partners.\u00a0Those partners are\u00a0biotech\u00a0and\u00a0pharma\u00a0companies\u00a0that are\u00a0currently trying out\u00a0their\u00a0technology.\u00a0\u00a0<\/p>\n
On Wednesday,\u00a0the\u00a0company came out of stealth with\u00a0$13 million\u00a0of funding, including pre-seed money, and a\u00a0seed round\u00a0led by\u00a0Fifty Years.\u00a0\u00a0<\/p>\n
\u201cCurrently, pharma and biotech companies either build entire teams of chemists\u00a0in-house\u00a0or work with contract research organizations overseas,\u201d\u00a0Tryin\u00a0said\u00a0of\u00a0the\u00a0process for\u00a0molecular synthesis. Human chemists\u00a0can spend\u00a0months of research to\u00a0create\u00a0even a single compound,\u00a0at a\u00a0cost\u00a0of\u00a0thousands of dollars.\u00a0\u00a0<\/p>\n
It involves a lot of trial and error\u00a0\u2014 studying various compounds, collecting data\u00a0on\u00a0biological activity, how\u00a0the\u00a0drug moves through\u00a0the\u00a0body, toxicology reports, and\u00a0coming up with\u00a0what to experiment with next.\u00a0\u201cThe\u00a0main limiting factor here is not testing\u00a0these compounds, but making\u00a0them in\u00a0the\u00a0first place,\u201d\u00a0Tyrin\u00a0continued. \u201cWe aim to compress this down to days.\u201d\u00a0\u00a0<\/p>\n
Tyrin\u00a0said\u00a0the\u00a0product is pretty straightforward.\u00a0Onepot\u00a0has a catalogue of molecules it can make. Clients choose which compounds\u00a0it want\u00a0and\u00a0then\u00a0Onepot\u2019s\u00a0technology will synthesize\u00a0the\u00a0molecules\u00a0and\u00a0then\u00a0ship\u00a0them\u00a0to\u00a0the\u00a0customer\u00a0so\u00a0the\u00a0customer can use\u00a0them in\u00a0their own experiments.\u00a0(They ship physical products either as dry compounds of solutions\u00a0in\u00a0plates or\u00a0vials).\u00a0\u00a0<\/p>\n
The\u00a0backend of\u00a0the\u00a0product is where Boiko and Tyrin have fun, dissecting\u00a0the\u00a0problems of chemical synthesis to find out\u00a0which combinations of molecules work together.\u00a0They built\u00a0a lab where\u00a0they are letting LLM agents access\u00a0these so-called molecule recipes\u00a0for\u00a0training\u00a0so\u00a0the\u00a0agents can also\u00a0find out what works and\u00a0doesn\u2019t\u00a0in compound building.\u00a0\u00a0<\/p>\n
\u201cWhen executing experiments in\u00a0the\u00a0lab, we capture every single detail that goes into\u00a0the\u00a0process,\u201d\u00a0Tyrin\u00a0said\u00a0\u2014 that means tracing temperature and, essentially,\u00a0the\u00a0ingredients that were added to a mixture to create compounds. \u201cNo information is lost, which makes experiences reproducible even if someone decides to run\u00a0them in ten years from now.\u201d\u00a0\u00a0<\/p>\n
This\u00a0also\u00a0means\u00a0their\u00a0agents generate\u00a0hypotheses\u00a0from real-world experiments rather than\u00a0literature data, often mined from\u00a0the internet.\u00a0\u00a0<\/p>\n
Boiko\u00a0called\u00a0the\u00a0fundraising process \u201chectic,\u201d and said\u00a0they met\u00a0their lead\u00a0investor\u00a0through an intro. \u201cWhat was supposed to be a short meeting turned into a multi-hour\u00a0whiteboard session about industrializing synthesis,\u201d\u00a0Boiko\u00a0said. Others in\u00a0the\u00a0round include Khosla Ventures,\u00a0Speedinvest, OpenAI co-founder Wojciech\u00a0Zaremba\u00a0and Google\u2019s Chief Scientist Jeff Dean.\u00a0\u00a0<\/p>\n
The\u00a0fresh capital will be used to build a second lab in San Francisco\u00a0so\u00a0the\u00a0team can take on more customers. It will also expand\u00a0the\u00a0team and its compound discovery engine.\u00a0On\u00a0the\u00a0service side, Boiko and Tyrin look at\u00a0WuXi\u00a0AppTec and Enamine as competitors.\u00a0\u00a0<\/p>\n
Overall,\u00a0Boiko and Tyrin hope to make drug discovery at least two times faster and hopefully change\u00a0the\u00a0perception\u00a0of what is possible after one taps into\u00a0the\u00a0\u201cweird\u201d chemistry that scientists once thought was off limits.\u00a0\u00a0<\/p>\n
\u201cYou\u2019re not just speeding up drug discovery, you\u2019re expanding\u00a0the\u00a0design space for what drugs and materials\u00a0can be,\u201d\u00a0Boiko\u00a0said. \u201cThat drug that we\u00a0haven\u2019t\u00a0discovered yet, might be out\u00a0there, waiting for us to find it.\u201d\u00a0\u00a0<\/p>\n","protected":false},"excerpt":{"rendered":"For Daniil Boiko\u00a0and\u00a0Andrei Tyrin,\u00a0the\u00a0idea for\u00a0Onepot\u00a0AI came from\u00a0the\u00a0same frustration.\u00a0\u00a0 \u201cThe\u00a0best ideas in drug discovery were often blocked not by…\n","protected":false},"author":2,"featured_media":148560,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[220,218,219,3749,61,60,1884,80,8976],"class_list":{"0":"post-148559","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-artificial-intelligence","8":"tag-ai","9":"tag-artificial-intelligence","10":"tag-artificialintelligence","11":"tag-chemistry","12":"tag-ie","13":"tag-ireland","14":"tag-startups","15":"tag-technology","16":"tag-venture"},"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/posts\/148559","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/comments?post=148559"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/posts\/148559\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/media\/148560"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/media?parent=148559"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/categories?post=148559"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/tags?post=148559"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}