Warp Sciences aims to industrialize laboratory chemistry with AI
Warp Sciences emerges with $215 million to build AI- and robot-assisted chemistry labs in the United States. The startup intends to sell compounds to biotech and pharma companies without developing its own molecule.
Warp Sciences, a startup founded by former executives from OpenAI, Meta, Terremoto Biosciences, and Devoted Health, has raised $215 million to build chemistry labs in the United States. Its idea is simple to state but harder to execute: produce, test, then learn from each experiment to accelerate drug discovery. The company positions itself as a CRO, that is, a contract research organization, combining chemists, AI, and robotics in the same loop.
From the promise of a model to the reality of a lab bench
Warp starts with a very concrete activity: optimizing molecules by synthesizing analogues to provide R&D teams with comparable variants in their search for a drug candidate. Kevin Weil, CEO and former chief product officer at OpenAI, says AI has accelerated software and mathematics, but not yet physical sciences at the same level. Here, the strategy is to close that gap where it really matters: in the lab, not in the brochure.
The startup claims to target small molecule programs in the preclinical phase, with compounds produced at milligram to multi-gram scale and typically long syntheses of five to fifteen steps. Clients retain control over their program and choose which molecules to pursue; Warp itself does not develop a proprietary pipeline. This commercial choice matters because it avoids the classic conflict between a service platform and a lab that would also want to act as a drug developer.
A model designed so that each experiment serves the next
The technological bet relies on a feedback loop between real experiments and models. Teams carry out syntheses, record results, then reinject this data to guide subsequent tests and further automate the work. Warp says it wants to keep both successes and failures, as a failed test can help define a method or detect an execution problem. Not glamorous, but very useful.
Sol Reisberg, co-founder and chief scientific officer, explains that the company wants to offer drug hunters chemistry of the same level as abroad, with intellectual property protection at the center of the relationship and without the premium pricing of high-end US CROs. According to the startup, the challenge is not only speed but also control over where sensitive work takes place. In this context, automation is not just about going faster: it must also make the chain more predictable and traceable.
Who is investing and who is steering the whole?
The funding round was led by Lightspeed, with participation from IVP, Mayfield, Lux, Premji Invest, Menlo Ventures, Conviction, and 01 Advisors, among others. Individual investors include Brian Armstrong, Ion Stoica, Dylan Field, as well as OpenAI members like Mark Chen and Sarah Friar [to be verified for the full list according to sources]. On the founding team side, DJ Patil brings operations and data experience, Mano Paluri expertise in advanced AI at Meta, while Sol Reisberg comes from medicinal chemistry.
Warp says it is already working with a small number of design partners, in weekly exchanges between its teams and those of its clients. This pace is telling: before promising an autonomous factory, it is necessary to show that an AI-driven chemistry service is reliable, delivers compounds on time, and does not break programs. The real battle will therefore be in execution, not on slides.
What this funding means for biotech
The $215 million raise gives Warp the means to build both physical capacities and software systems. The company says it charges per compound delivered, with a pricing agreement concluded with the client; if a compound is not produced, it is not billed. This model aligns revenue with work done, which is clearer than a time-based invoice and less likely to upset buyers. In chemistry, the best buzzword remains a usable result.
For biotech and pharma, the interest is clear: a US-based provider, better-controlled data, and the prospect of labs that learn from their own experiments. One very concrete question remains: will Warp be able to demonstrate measurable gains in time, quality, and cost without sacrificing scientific rigor? At this stage, the company announces the direction, not yet definitive proof.
Key points
- $215 million raised, according to sources.
- Lightspeed led the funding round.
- Warp is already working with some design partners.
- The service targets preclinical small molecules.
- The company does not develop its own pipeline.
- The model charges per compound delivered.
By the numbers
- $215 million — amount raised by Warp Sciences, according to sources.
- 5 to 15 steps — typical complexity of targeted syntheses, according to the startup.
- Milligram to multi-gram — announced production scale for compounds.
- United States — location of future chemistry facilities.