OpenAI’s Miles Wang targets $2B valuation for new AI drug discovery startup

This story is less about one researcher leaving OpenAI and more about a pattern: top frontier‑lab talent spinning out to build application‑layer companies that assume baseline foundation models are “good enough.” Wang is explicitly aiming AI at a domain where there is already rich structured data and an obvious commercial bottleneck: rediscovering viable uses for drugs that have already cleared basic safety hurdles. That’s a capital‑efficient thesis and exactly the kind of thing VCs want as AI funding shifts from model training to domain‑specific moats.

For the race to AGI, the move underscores how value will increasingly accumulate around problem‑specific workflows built on top of general‑purpose models rather than at the model layer alone. If a $2 billion valuation for a pre‑product spin‑out becomes normal for ex‑OpenAI talent, expect an explosion of specialist AI biotechs trying to own slices of the drug discovery stack before AGI arrives. Those companies won’t drive the frontier themselves, but they will generate domain data, safety regimes and market pull that feed back into what frontier labs prioritize.

It also hints at an internal equilibrium at OpenAI: once you believe scaling laws are mapped out, the career upside shifts to building category‑defining applications. That migration of talent may matter as much as any single architecture tweak in determining who captures AGI’s economic upside.