Miles Wang, an OpenAI researcher focused on applying AI to scientific and biological discovery, is leaving the ChatGPT maker to launch an AI drug discovery startup and is in talks to raise about $200 million at a $2 billion valuation. Lightspeed is in discussions to lead the round, and several other OpenAI researchers are expected to follow Wang out the door. The deal is not final and details could shift; Wang disputed the reported figures and company description but did not offer corrected numbers.
The valuation would put Wang's still-unnamed venture in the same tier as the current crop of AI-for-biology startups before it has publicly shipped anything. That's the market right now — investor appetite for AI drug discovery has decoupled from traditional biotech milestones, with credentials and model quality drawing pre-product term sheets in the hundreds of millions.
Wang's company may build AI models aimed at finding new uses for existing drugs, including compounds that previously failed in clinical trials, according to sources familiar with the plans. Repurposing FDA-approved molecules is a different bet than de novo drug design: safety data already exists, regulatory pathways are shorter, and time-to-revenue collapses from a decade-plus to something a venture-backed company can realistically underwrite.
Key facts
- 01Miles Wang is in talks to raise ~$200M at a $2B valuation for a new AI drug discovery startup, with Lightspeed leading discussions.
- 02Several other OpenAI researchers are expected to join Wang's new company, which may focus on finding new uses for existing FDA-approved drugs.
- 03Chai Discovery announced a $400M raise at a $3.8B valuation this week; Isomorphic Labs raised a $2.1B Series B in May.
- 04Wang joined OpenAI in 2024 after dropping out of Harvard's computer science program, and co-authored research on AI-driven scientific discovery.
The competitive set is filling out fast. Chai Discovery, a two-year-old startup building AI models that predict molecular interactions, announced on Tuesday that it raised $400 million at a $3.8 billion valuation. Chai co-founder Josh Meier is himself a former OpenAI researcher, making Wang's move the second high-profile OpenAI-to-biology jump in as many years.
Google DeepMind spinout Isomorphic Labs raised a $2.1 billion Series B in May, an order of magnitude larger than Wang's reported round and a signal of how much capital the category can absorb. Isomorphic is pursuing new drug candidates rather than repurposing; if Wang's team leans into the repurposing thesis, the two companies would be building adjacent rather than directly overlapping products.
Wang joined OpenAI in 2024 after dropping out of Harvard, where he had been working on a bachelor's degree in computer science. At OpenAI he co-authored research papers on evaluating how AI models can automate and accelerate scientific discovery — work that reads, in retrospect, like the intellectual foundation for a company like the one he is now raising for.
The broader talent flow out of frontier labs into vertical AI companies has become one of the defining patterns of this cycle. Researchers who built general-purpose reasoning and scientific-evaluation systems inside OpenAI, Anthropic, and DeepMind are increasingly deciding that the highest-leverage application of those systems is a focused company solving one industry's hardest problem, not another turn of the general-model crank.
Drug discovery is a natural target. The economics are enormous — a single repurposed compound reaching approval can generate billions in revenue — and the technical bottleneck is exactly the kind of prediction problem large models have started to solve. Model quality on protein structure, molecular interaction, and clinical outcome prediction has improved sharply in the last 24 months, though translating benchmark gains into approved medicines remains unproven at scale.
The risks worth flagging: no AI drug discovery company has yet delivered an approved, AI-designed drug to market, and repurposing bets depend on the FDA accepting model-generated hypotheses as sufficient justification for new indication trials. Wang's team will also be competing for a scarce pool of ML researchers who understand both frontier models and biology, and payroll at that intersection is not cheap. A $200 million round buys a lot of compute, but the road to a data readout is measured in years, not quarters.
The deeper story is what Wang's departure says about where the leverage sits in AI right now. When researchers who could name their price at OpenAI are instead raising two-comma seed-adjacent rounds to attack a single industry, the implicit bet is that vertical AI companies with strong founding teams will out-earn generalist model providers on the problems that actually matter. If Wang and his co-founders ship a repurposed drug through Phase 2 in the next few years, that thesis stops being speculative — and the flow of talent out of frontier labs becomes a torrent.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
