Miles Wang, a researcher at OpenAI whose work has centered on applying AI to accelerate biological and scientific discovery, is reportedly departing the ChatGPT maker to found a drug discovery startup—and he's not leaving alone. Several OpenAI colleagues are expected to follow him out the door.

According to sources familiar with the situation, Wang is deep in talks to raise roughly $200 million at a $2 billion valuation, with Lightspeed reportedly in discussions to lead the round. Wang himself pushed back on those figures when contacted, though he declined to offer corrections. Lightspeed stayed quiet. So we're in the standard pre-announcement fog where everyone knows something is happening but nobody wants to say exactly what. Classic.

What the Startup Might Actually Do

Here's where it gets interesting. Rather than chasing the crowded field of de novo drug design—building novel molecules from scratch, which is expensive, slow, and has a brutal failure rate—Wang's venture may focus on drug repurposing. That means using AI models to identify new therapeutic uses for drugs that are already FDA-approved, or potentially resurrecting compounds that flamed out in clinical trials for one indication but might work for another.

This is a genuinely smart wedge. Repurposing an approved drug sidesteps the most time-consuming part of the development pipeline: safety validation. You already know the compound won't kill the patient in the ways regulators care about. If your AI can match an existing molecule to an unmet clinical need with precision, you can compress a decade-long development timeline into something far more investable. The tradeoff? The low-hanging fruit in repurposing has been getting picked for years, and you're competing against decades of human pharmacological intuition. The AI better be doing something genuinely novel, not just running fancier literature searches.

The Funding Climate Is Clearly Losing Its Mind (In a Good Way?)

Wang's raise—if it closes at the reported figures—would drop into a sector that is suddenly drowning in capital. On the same day this story broke, Chai Discovery, a two-year-old startup building AI models for predicting molecular interactions, announced a $400 million raise at a $3.8 billion valuation. Its co-founder Josh Meier is also an OpenAI alum, which is either a remarkable coincidence or a sign that OpenAI's biology-adjacent research teams have become the hottest talent pipeline in biotech. Probably both.

Meanwhile, Isomorphic Labs—the Google DeepMind spinout doing AI-driven drug discovery—closed a $2.1 billion Series B in May. When DeepMind alumni and OpenAI alumni are both raising multi-billion-dollar rounds in the same quarter, it's fair to say the "AI for drug discovery" thesis has officially graduated from speculative to consensus. Whether that consensus is correct, or whether we're building toward a graveyard of expensive clinical trial failures, is a question the market won't answer for years.

The Founder Backstory

Wang joined OpenAI in 2024 after leaving Harvard mid-degree, where he was pursuing computer science. He co-authored research on automating and accelerating scientific workflows in wet labs—the kind of painstaking, pipette-and-centrifuge experimental work that is both essential and brutally slow. The pitch that AI can handle more of that cognitive and logistical load is credible; OpenAI's own published work in this area showed measurable gains in research throughput.

The Harvard dropout credential, once a punchline, has apparently become a legitimate signal to investors again. Whether that says something profound about the current opportunity in AI or something uncomfortable about how venture capital evaluates risk is left as an exercise for the reader.

What to Watch For

• The actual technical differentiation. Drug repurposing via AI isn't new—startups have been pitching it for the better part of a decade. What matters is whether Wang's models can surface candidates that human researchers and existing tools genuinely miss, and whether those candidates survive clinical validation.
• The OpenAI talent pipeline effect. Two major biotech raises in one week from OpenAI alumni suggests either a coordinated exodus or a cohort of researchers who've been quietly planning their exits. Either way, OpenAI's biology research team is worth watching.
• Lightspeed's portfolio construction. If the VC giant is leading this round alongside its other life sciences bets, it tells you something about how top-tier funds are positioning for the next phase of AI applications beyond software.

The $2 billion valuation on a company that hasn't yet announced its name, product, or team publicly is the kind of number that makes engineers raise an eyebrow. But if the underlying science is sound and the repurposing thesis lands even one major clinical win, the math starts to look a lot more reasonable. Drug development is a brutal, expensive game—but the prizes are enormous and the incumbents are slow. That's the gap AI is trying to walk through.