OpenAI researcher is building a $2B AI drug discovery startup

Signals Inbox·July 15, 2026·Biopharma

OpenAI researcher Miles Wang is reportedly leaving to build an AI drug discovery startup and discussing a $200 million raise at a $2 billion valuation. The tension is hard to miss: the company reportedly has no public name or product yet, but investors may already price it like an established biotech unicorn.

Q1What actually happened?

According to TechCrunch’s original report, Miles Wang is leaving OpenAI to start a company building AI models for drug discovery. The startup is reportedly discussing a roughly $200 million round at a $2 billion valuation, with Lightspeed in talks to lead. The terms are not final, and Wang has disputed parts of the report without publicly giving corrected numbers.

Q2Why is the $2 billion valuation so unusual?

Because the company reportedly has no public name, launched product, pharma customer, or clinical drug candidate yet. Investors would mainly be buying the team and its potential. That is the same founder-first logic seen in frontier AI labs, now moving into biotech, where proving that a product works normally takes years.

Q3What would the company actually build?

The reported plan is to train AI models that find promising drug opportunities. One possible focus is drug repurposing: finding new uses for medicines that already exist or reconsidering drugs that failed earlier trials. That can be faster than inventing a molecule from zero because some safety and manufacturing work may already exist, but the drug still needs real biological and clinical proof.

Q4How does this compare with existing AI drug startups?

The bar is already high. Chai Discovery announced a $400 million round at a reported $3.8 billion valuation the same week, after building molecular design models used by major pharmaceutical companies. Isomorphic Labs raised $2.1 billion in May 2026 and is backed by Alphabet and years of AlphaFold research. Wang’s startup could start near their valuation range before showing comparable commercial or scientific proof.

Q5Why are investors willing to move this early?

Top AI researchers have become scarce assets. Investors are betting that Wang can recruit other OpenAI scientists, train powerful scientific models, and attract pharmaceutical partners before competitors do. The valuation is therefore less a judgment on an existing business and more an expensive attempt to secure a potentially elite research team early.

Q6Has AI already proven it can create successful drugs?

Not at the level these valuations suggest. AI is clearly useful for protein prediction, molecule design, screening, and lab planning, and more AI-assisted drugs are entering human trials. But clinical trials remain the hard part. Human biology can still kill a promising candidate after years of work, and the field has not yet produced enough approved drugs to prove that AI has rewritten the economics of pharma.

Q7So what is the real signal?

AI drug discovery is beginning to inherit frontier AI’s founder premium. A small group of researchers can now discuss billion-dollar valuations before launching a product because investors fear missing the next major scientific platform. The number to watch is not the valuation. It is how quickly the company produces a working model, a serious pharma deal, and eventually a drug that survives clinical trials.