Business & Deals / Healthcare

OpenAI’s Bet on Faster Drug Development

AI’s real money in pharma may be in speeding up clinical trials, not discovering drugs, and OpenAI wants a cut of the time it saves.

OpenAI’s Bet on Faster Drug Development

On October 20, OpenAI announced its acquisition of the biotech company Orchestra Bio. The exact financials of the deal were not disclosed. Orchestra was founded in 2022, and since then it has raised around $20 million in investment. By OpenAI’s standards, this deal may seem small, but it is very telling about how AI companies view the future of the biotech and pharmaceuticals industries. There has been lots of excitement in the biotech industry around AI, but most of it has been around drug discovery. Orchestra works on what comes next, with the drug studies, vendors, and deadlines being the major steps between a promising molecule turning into an approved drug. This acquisition shows the breadth of use AI has in the industry, rather than just drug discovery, which people have most focused on.

Orchestra’s platform lets scientists and program managers hand off large parts of monitoring their drug programs to software and AI agents. The process from discovering a drug to getting it approved is a long and complex process. Orchestra’s technology hopes to simplify it and make it more efficient. One example OpenAI gives is that if a biotech company has a failed manufacturing batch, it can affect other activities and timelines. In that situation, Orchestra’s agents would gather all relevant information and present updated solutions and timelines to the scientists and program managers, who can then make a more informed decision. OpenAI is integrating Orchestra’s team into Rosalind, which is its AI tool for life sciences research.

Developing a drug from scratch takes roughly 10-15 years, yet most of the AI investment that has gone into the pharmaceutical industry has gone into the very start of the process, molecule design. This means far less attention has gone into clinical trials, which are the longest stage of development. A single day of delay of the launch of a drug can cause hundreds of thousands of dollars to be lost in sales, and the running of clinical trials an extra day can cost tens of thousands of dollars. In the majority of cases, drug development does get delayed, which shows the value of making the process more efficient. Cutting just a small percentage of the 10-15 years of development could save the company millions of dollars. Speed matters even more because a drug’s twenty-year patent clock starts at filing, which is years before it is ever sold. Cutting down development could be worth tens of millions of dollars in earlier sales.

OpenAI, with its Rosalind division, isn’t the only AI company moving into pharma. In April, Anthropic bought the startup Coefficient Bio for $400 million. The startup had fewer than ten employees at the time they bought it, meaning they paid over $40 million per employee. This shows how much AI companies are willing to pay to get into the industry. Anthropic also has its own Claude for Life Sciences division, which they integrated the Coefficient Bio team into. It has a large focus on assisting in drug discovery. The fact that two of the biggest AI companies are investing in the industry indicates the potential value AI will have.

The big question is how these companies plan to monetize their AI platforms and agents. Currently, OpenAI and Anthropic mostly sell subscriptions and charge usage-based fees. But earlier this year, OpenAI said that as they move into different specialized fields like drug discovery and biotech, new business models like licensing and outcome-based pricing will emerge. In other words, OpenAI plans to charge a fee that gives them a percentage of the value created from their platform. With use in drug discovery, it could be very difficult to determine the amount of value created using AI tools. However, in the case of Orchestra Bio, it's much easier to tell how much value is gained by looking at how much faster the drug achieves and passes its trials.

Overall, the thing to watch is not whether Orchestra’s software works, but whether OpenAI begins to sign contracts that pay it based on value created, or in this case, time saved. If it does, AI’s push into pharma looks less like selling software and more like taking stakes in the medicines themselves. In an industry like biotech where each day costs hundreds of thousands, AI has enormous potential to turn saved time into a valuable product.