Kevin Weil launches Warp Sciences with $215 million for robotic drug-discovery labs

Former OpenAI science lead Kevin Weil launched Warp Sciences on October 22, 2026, with $215 million in funding to build robotic labs for drug discovery. The startup aims to make and test compounds for pharmaceutical and biotechnology companies, pairing a U.S.-based chemistry service with a longer-term effort to train AI on real experiments.

The financing values Warp at $865 million, according to Weil’s announcement as reported by TechCrunch. Lightspeed led the Series A, with IVP, Premji Invest and Lux Capital also participating.

Weil is CEO. His co-founders include chief operating officer DJ Patil, the first U.S. chief data scientist under President Obama; chief scientific officer Sol Reisberg, a medicinal chemist; and Mano Paluri, who leads models and robotics. Paluri spent 14 years at Meta and was most recently a vice president at Meta Superintelligence Labs.

Chemists first, robots over time

Warp’s initial service is lead optimization: refining a promising molecule before it enters human trials. Chemists design variations, then make, purify and test them. The results guide the next round. According to Lightspeed, good contract laboratories aim to complete each round in two to three weeks; a full program usually takes two to four years.

Lightspeed says expert chemists will initially perform much of the work, using AI tools to plan how to make compounds and select experiments. Robots are intended to take over repetitive tasks over time, including setting up reactions and purifying samples. Chemists would focus on harder problems.

Part of that training will involve human chemists wearing cameras and sensor gloves to demonstrate laboratory tasks, TechCrunch reported. Warp’s goal is robots that can work around the clock.

Keeping the experiments that fail

In the IVP interview, Weil argued that chemistry and biology need experimental data at a scale that can support better scientific models. Lightspeed describes Warp’s system as recording every step of every run, including failures, alongside chemistry results and data about robot operations and laboratory processes.

The investor argues that published chemistry data skews toward successes. Warp intends to use its records to identify slow or unreliable laboratory steps and train models to predict how to make compounds and which ones are worth making. Lightspeed says the company later plans to sell predictions as well as compounds.

Selling a U.S. alternative

Location is also part of the pitch. Weil wants to bring outsourced drug-discovery chemistry back to the United States and reduce geopolitical supply-chain risks. Biotech companies have long sent labor-intensive molecule synthesis to contractors in China, including WuXi AppTec. Weil says Warp will focus on its initial service before considering other scientific industries.