Biossil Raises $153M in OpenAI-Led Round to Push Repurposed Failed Drugs Into Trials
核心洞察
Biossil Inc. (搜索) has raised US$153 million in an all-equity financing led by OpenAI (搜索)'s Startup Fund, valuing the Toronto drug developer at US$1 billion.
The company uses OpenAI (搜索) large language models to identify shelved drug candidates that failed late-stage trials and revive them for new indications or defined patient groups.
Biossil has bought or licensed 12 molecules, including a precision antibiotic from Summit Therapeutics for US$500,000 up front.
OpenAI (搜索) has led a US$153-million all-equity financing of Biossil Inc. (搜索), the Toronto startup that applies artificial intelligence to failed drug candidates in an effort to return them to clinical development. The AI company provided the majority of the round through its OpenAI Startup Fund, and the deal values Biossil at US$1 billion.
The financing was not publicly announced; details were shared by a source familiar with the matter who was not authorized to discuss the information. Biossil chief executive Anthony Mouchantaf declined to comment. Founders Fund (搜索), which led a previous US$22-million financing of Biossil in 2024, participated in the new round alongside existing investors Quiet Capital, Modern Capital, Golden Ventures and Panache Ventures, plus Duke University's endowment fund and sovereign wealth fund Abu Dhabi Investment Council. OpenAI (搜索) co-led Biossil's previous venture financing a year ago with Founders Fund, when the company raised US$43 million.
According to the source, Biossil will use the proceeds to expand its capacity for clinical development and to selectively acquire more drug candidates. A small portion of the all-equity deal is being paid to early employees. In an August filing with Corporations Canada, the company disclosed it had created a new class of 5.37 million preferred shares valued at US$28.69 each, triple the value of shares issued in last year's financing.
A Platform Built on Salvaged Molecules
Biossil, co-founded in 2023 by Mouchantaf and Alexander Mosa, built a software platform using OpenAI (搜索)'s large language models to identify promising molecules from what the company describes as the pharmaceutical discard pile. Its algorithms search through drug candidates that failed trials at advanced stages, looking for ways to give them a second chance. Some drugs failed because they worked only in certain patients with particular attributes; in other cases, trials were poorly designed or pursued the wrong outcomes.
The founders reasoned that if the molecules could be brought back into trials in a modified fashion, they could forgo years of studies and hundreds of millions of dollars in costs on the path toward regulatory approval. Several companies have designed drug candidates using AI and taken them into trials; none have yet reached the market.
Biossil has bought or licensed 12 molecules. In July it paid US$500,000 up front for a precision antibiotic from Summit Therapeutics Inc., with additional payments contingent on regulatory and commercial milestones and market entry. That drug faltered in a late-stage efficacy trial five years ago as a treatment for patients with C. difficile infections.
Sickle Cell and Glioblastoma Programs Advance
Two Biossil molecules target sickle cell disease (搜索), a genetic condition that predominantly affects people of African or Indian descent. Johnson & Johnson and Pfizer Inc. each attempted to develop the drugs with biotechnology startups but abandoned those programs after they failed late-stage efficacy trials. In both cases, Biossil's AI uncovered reasons to revisit the candidates.
For the J&J drug, senicapoc, the trial failed to show significant improvement over placebo in relieving pain, the study's main goal. The data did show the drug was better at preventing the breakdown of red blood cells, which causes anemia and other serious complications. Biossil is now sponsoring a sickle cell disease (搜索) trial with the abandoned J&J molecule, with recruiting under way at Toronto's Hospital for Sick Children (搜索) (SickKids), Centre hospitalier de l'Université de Montréal and McGill University.
The company is also launching a trial in Germany using the same molecule to treat glioblastoma (搜索), an aggressive cancerous brain tumour, in partnership with Heidelberg University Hospital and the German Cancer Research Center, and a separate trial for the same indication in Denmark with Aarhus University.
Academic and Hospital Partnerships
Biossil has partnered with universities and research hospitals in Canada, the United States and Europe, including SickKids, Harvard University and the Mayo Clinic. It recently signed trial collaboration agreements with SickKids, Toronto's University Health Network and Stanford University's Skin Innovation and Interventional Research Group.
The financing comes as OpenAI (搜索) and rival Anthropic (搜索) expand beyond broadly used AI models into sector-specific tools, including health care. Earlier this month, Anthropic said it had established a wet laboratory in the San Francisco Bay Area staffed by a life-sciences research group using its Claude AI assistant, and said Claude had already "discovered a novel enzyme system with properties reminiscent of CRISPR," a genetic modification technology used by medical researchers.
For practitioners tracking AI in drug development, the relevant question is not a model release but whether a financing-backed attempt to apply AI to a constrained clinical-development workflow produces new treatments. Evidence from later trials and regulators, not the funding announcement, will determine that outcome.