Chemical labs under robotic control
Kevin Weil, former Science Lead at OpenAI, announced on Tuesday the emergence from stealth mode of Warp Sciences, a US$ 215 million startup built to construct the infrastructure that will enable biotech and pharmaceutical companies to discover new medicines faster. The company, valued at US$ 865 million post-money, merges medicinal chemistry, artificial intelligence, and robotics in a model designed to tackle one of the oldest bottlenecks in drug discovery: the need for manual, repetitive synthesis of molecules.
Warp Sciences won't develop drugs on its own. The focus, as co-founder and Chief Operating Officer DJ Patil — the first Chief Data Scientist of the White House — made clear during the announcement, is to build the infrastructure for drug developers. "Our biotech and pharma customers determine which molecules to pursue and retain control of their drug programs," Patil said at the launch, reinforcing that the startup operates as a service partner rather than a competitor.
What the startup will do
Warp's initial service focuses on lead optimization — the stage at which researchers have already identified a promising molecule and must create and test hundreds or thousands of variations of that molecule to improve its activity, reduce side effects, or refine other properties before advancing the candidate into clinical development. It is one of the most labor-intensive and time-consuming steps in the pharmaceutical pipeline.
For years, this part of the work has been outsourced to contractors in China, where companies like WuXi AppTec compete on speed and cost. With trade tensions and recent legislation like the BIOSECURE Act, signed in December 2025, many companies have been seeking domestic alternatives. Warp aims to repatriate part of that capacity to the United States, combining expert human chemists with robots and AI models trained on physical experiment data.
Training robots with expert chemists
Warp's most unique approach is its robot training method: expert chemists wear cameras and pressure-sensing gloves to record every movement performed in the lab. These visual and tactile data feed the startup's AI models, which learn to execute complex laboratory tasks — from reagent measurement to sample handling — autonomously.
"AI has dramatically accelerated software engineering and mathematics, but it has yet to make the same impact in the physical sciences," Weil told reporters. "Chemistry and biology move forward when you make something, test it, and learn from the result. If we want to build more capable AI for science, we need to generate real-world data at scale."
The plan is for the robots to operate 24 hours a day, seven days a week, eliminating the natural breaks of human researchers and dramatically accelerating the hypothesis-experiment-response cycle. With larger volumes of experimental data, Warp's own AI models become more accurate, which makes the robots more efficient, generating a virtuous cycle of continuous learning.
The team behind it
Warp Sciences assembles an unusual roster of heavy-hitting names. Weil, who most recently served as Chief Product Officer and Vice-President of Science at OpenAI, co-founded the company alongside Solomon "Sol" Reisberg — a medicinal chemistry leader at Terremoto Biosciences, an AI-driven drug discovery startup; Manohar "Mano" Paluri, former Vice-President of Meta Superintelligence Labs and founding member of FAIR (Fundamental AI Research); and DJ Patil, former executive at Devoted Health and the first U.S. Presidential Chief Data Scientist.
The angel investor list is equally formidable: Sarah Friar (former OpenAI CFO), Reid Hoffman (LinkedIn co-founder), Brian Armstrong (Coinbase CEO), Mike Krieger (Instagram co-creator and current Anthropic Labs co-lead), Dylan Field (Figma CEO), alongside names from the traditional pharmaceutical industry.
The market and the challenges
The global contract research organization (CRO) services market already turns over tens of billions of dollars, and the chemical synthesis segment is one of the most profitable. Warp enters a landscape dominated by established Asian providers with decades of accumulated scale and know-how. The competitive advantage proposed by the startup is the combination of domestic capacity, robotic automation, and AI-driven learning — something no current competitor offers in an integrated fashion.
The risk, however, is significant. Robots in wet chemistry labs must deal with the unpredictability of real reactions, and the transition from a pilot laboratory to large-scale operation has never been trivial. Warp has already acquired robotic arms and AI chips but still has around 20 employees — a small number compared to the thousands of people operating the largest synthesis centers in China.
The promise of "every experiment making the next one smarter" will be tested not just by speed, but by the reliability of the generated data. For biotech companies, the quality of experimental data is everything: a poorly executed synthesis can unfairly discard a promising candidate, or worse, advance with undetected flaws. The standard of reproducibility that Warp must achieve to gain traction is extremely high.
What to watch
Warp Sciences marks the most ambitious frontier yet at the intersection of generative AI and the physical sciences. While other startups focus on computational protein modeling or molecular structure prediction, Warp bets that the next leap lies outside the computer — inside the laboratory. If it can scale its operation and demonstrate that its robots can reliably reproduce the most complex chemical protocols, it could reshape not only the economics of drug discovery, but also the relationship between artificial intelligence and the physical world. Whether robotic automation of chemical labs will become the open-source revolution for the natural sciences remains to be seen.
Sources: TechCrunch, Yahoo Finance / Business Wire, Drug-Dev.com
✓ Independent sources cross-checked and verified before publishing
