Warp Sciences, founded by former OpenAI science lead Kevin Weil, raises $215M to automate drug-chemistry labs
The AMW Read
A new well-funded entrant led by ex-OpenAI staff that is explicitly built as a US alternative to Chinese contract research. It updates the AI-bio player map, and US-China decoupling is the core business rationale.
Warp Sciences, founded by former OpenAI science lead Kevin Weil, raises $215M to automate drug-chemistry labs
Warp Sciences (워프 사이언스) has launched with a $215 million Series A (about 280 billion won) at an $865 million post-money valuation, according to Bloomberg as cited by AI Times. Lightspeed Venture Partners led the round. IVP, Mayfield, Lux Capital, Premji Invest and Menlo Ventures also took part. Angel investors include OpenAI CFO Sarah Friar, former OpenAI executive Fidji Simo, Coinbase CEO Brian Armstrong, Palantir CTO Shyam Sankar and Reid Hoffman. Weil's co-founders are Manohar Paluri, a former vice president at Meta's superintelligence research group; DJ Patil, the first US chief data scientist; and Solomon Reisberg, a former executive at Terremoto Biosciences. The company has about 20 staff and has already bought robot arms and AI chips. It will not develop its own drugs. Instead it will sell lead-optimization services to pharma and biotech companies: synthesizing and testing hundreds to thousands of variants of a candidate molecule.
Warp is entering contract research, not AI drug discovery. Its target is work that pharma has long outsourced to Chinese providers such as WuXi AppTec, which compete on low cost and fast turnaround. Weil argues that US-China tension and fears about sharing intellectual property with Chinese contractors make that dependence a risk. He also says that bringing the work back to the US only works if automation can match Chinese cost. The plan starts with chemists doing much of the work while AI picks synthesis routes and next experiments. Chemists will wear cameras and pressure-sensing gloves, and that data will train models so robots can take over repetitive steps such as reaction setup and purification. Over time the goal is lab work running around the clock.
Investors should watch the data loop more than the valuation. The bet only pays off if data from chemists' physical lab work turns into robot skills faster than WuXi can cut prices. Lightspeed says it can steer its life-sciences portfolio companies to Warp as customers, so early revenue may come from inside the investor network. Weil also says he would like to use US-developed open-weight models alongside frontier models. That leaves model sourcing an open question for any team building AI-plus-robotics lab services.