OpenAI researcher Miles Wang in talks to launch AI drug discovery startup valued at $2B

Miles Wang, a researcher at OpenAI known for applying artificial intelligence to scientific and biological research problems, is preparing to depart the company to build a new venture centered on AI-driven drug discovery, according to four people familiar with the matter. The plans, still taking shape, would see Wang exit one of the world's most closely watched AI labs to pursue a narrower but increasingly crowded frontier: using machine learning models to speed up how new medicines are found and developed.

Wang would not be going it alone. Multiple people briefed on the matter say a handful of additional OpenAI researchers are expected to follow him into the new company, suggesting the effort is being built around a team with direct experience applying frontier AI techniques inside one of the industry's most resourced research organizations.

On the financial side, two people with knowledge of the discussions say Wang has been negotiating an investment of roughly $200 million that would value the still-forming company at about $2 billion — a striking figure for a startup that has not yet formally launched. Sources say venture firm Lightspeed is among the parties negotiating to lead the financing. Because the round remains under discussion, the eventual terms, participants and valuation could shift before anything is finalized.

Reached for comment, Wang pushed back on how the prospective deal was characterized, taking issue with the reported valuation and funding total as well as the broader description of the company, though he did not offer alternative figures or clarify what aspects were inaccurate. Lightspeed did not respond to a request for comment on the talks.

The move fits a pattern that has become increasingly common across the AI industry, where researchers with experience at major labs like OpenAI, Google DeepMind and Anthropic have spun out to start their own ventures, often commanding outsized valuations before shipping a product. Investors have shown particular appetite for teams applying large-scale machine learning to biology and chemistry, betting that techniques originally built for language and reasoning tasks can meaningfully shorten the years-long, capital-intensive process of identifying viable drug candidates.

That enthusiasm has already fueled a wave of well-funded entrants in AI-assisted biotech, though the sector's track record remains unproven at scale — promising computational predictions do not always translate into successful clinical outcomes, and drug development remains subject to lengthy regulatory review regardless of how a candidate compound was identified. A venture led by a research team pulled directly from OpenAI would nonetheless carry an unusual pedigree, potentially giving it access to techniques and talent networks not available to earlier entrants in the space.

Much about the venture remains undetermined publicly, including its name, structure and product roadmap, and Wang's disputing of key reported details underscores that the arrangement is not yet locked in. Whether the $200 million round closes at the reported valuation, who ultimately joins Wang's team, and how the company positions itself within an already competitive AI-for-biology field are all likely to become clearer only once formal fundraising concludes and Wang's departure from OpenAI is confirmed.

Source: TechCrunch