OpenAI has acquired Orchestra Bio, a startup developing AI agents for pharmaceutical drug development, in a move shared exclusively with Axios. The acquisition allows the ChatGPT maker to integrate Orchestra's team and technology into Rosalind, a suite of AI models and tools designed for life sciences researchers. Financial terms for the deal were not disclosed.
Orchestra Bio's platform enables scientists to delegate specific tasks to AI agents, including monitoring studies, tracking drug development timelines, and identifying potential delays. Both companies aim to deploy agents that connect scientific findings with the decisions and next steps required to advance potential treatments. This expansion reflects a broader trend where AI giants like OpenAI and Anthropic are pushing into science to extend their capabilities beyond core competencies in coding and cybersecurity.
OpenAI introduced the specialized GPT-Rosalind model in April to assist researchers with complex problems in biology, chemistry, and drug discovery. The company has also partnered with pharmaceutical giants such as Novo Nordisk and Moderna to integrate AI into research and development workflows. OpenAI states that developing a new medicine typically takes 10 to 15 years, a timeline it hopes AI can shorten.
The acquisition follows similar moves by rival Anthropic, which acquired AI biotech startup Coefficient Bio for roughly $400 million in April, according to The Information. Anthropic has continued its push into life sciences with the June launch of Claude Science, an AI-powered research platform connecting scientific databases, analysis tools, and specialized agents. In September, Anthropic announced a new biological research lab and reported that Claude helped discover a previously unknown enzyme system with similarities to CRISPR, though some scientists described these findings as overblown, per Bloomberg.
Several AI investors told Axios that a key question as AI companies go public is how far they can expand into new markets. The shift marks a transition from building the best models to applying intelligence across industries.