Ex-OpenAI science head launches Warp Sciences with $215M

Bringing drug labs home

Kevin Weil, the former head of OpenAI for Science, has officially stepped into the startup world with the launch of Warp Sciences. The company recently emerged from stealth, securing $215M in funding led by Lightspeed Venture Partners. This investment places the company's valuation at $865M.

The primary goal of Warp Sciences is to automate a specific, labor-intensive phase of drug development. For years, pharmaceutical companies have sent this work to overseas contractors, particularly in China. By building a fully automated lab within the United States, Warp Sciences intends to provide a domestic alternative for these critical tasks.

The move comes at a time of heightened tension regarding drug supply chains. Some drugmakers have expressed concerns about the risks of sharing sensitive research data with Chinese contractors. Additionally, some companies are looking to distance themselves from firms like WuXi AppTec, which has faced scrutiny from the US government over potential links to China's military.

The mechanics of automation

Warp Sciences does not intend to discover new drugs from scratch. Instead, it focuses on the refinement process. Once a researcher identifies a promising molecule, they must create hundreds or thousands of variations to improve effectiveness and reduce side effects. This repetitive, high-volume work is what Warp Sciences aims to automate.

To achieve this, the startup's team of about 20 people has already begun purchasing robotic arms and AI chips. The company is using a unique method to generate training data for its models. Chemists will wear cameras and pressure-sensing gloves while performing experiments. This setup allows the company to capture the nuances of human movement and touch to teach its AI how to replicate those actions in a robotic environment.

Weil noted that this approach is necessary because the field of drug development is heavily driven by physical experiments. He explained that frontier AI models are not yet changing the world in the way they have for software engineering or math because the models need to see specific experimental data to be effective.

A growing AI ecosystem

The launch of Warp Sciences is part of a broader trend of former OpenAI executives and researchers moving into the biotech space. Weil is the second former OpenAI leader this week to raise significant capital for US-based AI infrastructure. On Tuesday, Bob McGrew’s company, Arda, raised $232M to set up AI-driven factory lines.

Investors are increasingly pouring money into the intersection of AI and drug research. Other notable moves in this space include OpenAI-backed Chai Discovery raising $400M in July and OpenAI's recent acquisition of Orchestra Bio, a company that helps drugmakers run development with AI agents. Furthermore, OpenAI researcher Miles Wang was reported to be raising approximately $200M for his own drug startup.

The leadership team at Warp Sciences reflects this high-level expertise. Co-founders include Manohar Paluri, a former vice president at Meta’s Superintelligence Labs; DJ Patil, who served as the White House’s first chief data scientist; and Solomon Reisberg, who previously worked at Terremoto Biosciences. These figures represent a concentrated effort to apply high-level machine learning to the physical constraints of chemistry and biology.

What comes next

The success of Warp Sciences will depend on whether its automated lab can match the speed and cost-efficiency of the Chinese firms it seeks to replace. If the technology scales, it could significantly shorten the timeline for bringing new medications to market while keeping the data within US borders.

The company will likely focus on building a proprietary dataset of chemical interactions. By capturing the 'how' of chemistry through wearable sensors, they hope to create a moat of high-quality data that general-purpose models lack. This represents a shift toward 'embodied' AI, where the model's intelligence is tied directly to its ability to interact with the physical world.