AI-based drug discovery company Recursion raises $121M at a $646M valuation; VCs invested about $1B in AI drug discovery startups in 2018, up from $237M in 2016
Robert Langreth / Bloomberg :
Context & Ripple Effects
Recursion's $121M round at a $646M valuation lands in the middle of the first real funding wave for AI drug discovery: VCs put roughly $1B into the category in 2018, a fourfold jump from $237M in 2016, and London rival BenevolentAI had already raised $115M at a $2.1B post-money valuation months earlier. The Bloomberg piece captures a sector where valuations are set more by model ambition than by approved drugs.
The arc that follows validates the thesis and reshapes it: Recursion went on to raise a $239M Series D alongside a Bayer partnership for fibrotic diseases, then took $50M from Nvidia to train its models on Nvidia's platform against its own 23,000TB+ datasets — a shift from selling AI as a service to pharma toward owning compute-heavy infrastructure.
First-order effects
- Recursion gets a war chest to scale its drug-discovery pipeline at a moment when the category's 2018 fundraising (~$1B) has outpaced any approved drug, making capital efficiency the differentiator among AI-discovery startups.
Second-order effects
- Rivals like BenevolentAI, already valued at $2.1B versus Recursion's $646M, face pressure to convert valuations into pharma partnerships — the path Recursion later took with Bayer — or concede that headline valuations outrun revenue.
Third-order effects
- The pattern points toward consolidation around compute-backed platforms: Recursion's later Nvidia tie-up and Chai Discovery's 2026 $400M raise at a $3.8B valuation pitching models as AI infrastructure suggest the sector's endgame is infrastructure ownership, not per-program discovery fees.
The trend: AI drug discovery is evolving from data-science startups chasing pharma fees into compute-heavy infrastructure platforms, with chipmakers and mega-rounds replacing early-stage VC as the sector's capital base.