AI Boosts Drug Manufacturing: YC-Backed ReactWise Leads the Way

Artificial intelligence is shaking things up in the world of chemistry, and ReactWise, a Cambridge, U.K.-based startup backed by Y Combinator, is at the forefront. They're using AI to speed up chemical manufacturing, a crucial step in getting new drugs to market. Once a promising drug is identified in the lab, pharmaceutical companies need to produce larger quantities for clinical trials. This is where ReactWise steps in with its "AI copilot for chemical process optimization," which they claim can accelerate the standard trial-and-error process by 30 times. "Making drugs is a lot like cooking," explained co-founder and CEO Alexander Pomberger in a chat with TechCrunch. "You need to find the best recipe to make a drug with high purity and high yield." Traditionally, the industry has relied on trial-and-error or staff expertise for this "process development." But by adding automation, ReactWise aims to reduce the number of iterations needed to find the perfect recipe. The startup is working towards "one shot prediction," where the AI can predict the ideal experiment almost immediately, without the need for multiple iterations. Pomberger believes they'll achieve this within two years. Even now, their machine learning models can significantly cut down the number of iterations required, saving time and resources. Pomberger, a chemist by training with experience in Big Pharma, saw firsthand how tedious and trial-and-error driven the industry can be. ReactWise is essentially condensing five years of his academic research into what he calls "a simple software." This research focused on automating chemical synthesis using robotic workflows and AI. To build their AI models, ReactWise has conducted thousands of reactions in their labs, using a "high throughput screening" method to test 300 reactions at once. This speeds up the process of gathering training data. Pomberger noted that in pharma, a handful of reaction types are used repeatedly. By generating thousands of data points for these key reactions, they can train foundational reactivity models that fundamentally understand chemistry. This means when a client needs to develop a scalable process, they don't have to start from scratch. The startup began capturing these reaction types last August and aims to complete this by summer, targeting 20,000 chemical data points to cover the most important reactions. Traditionally, getting a single data point can take a chemist one to three days, making it an expensive and time-consuming process. So far, ReactWise has focused on manufacturing processes for small molecule drugs, which can be used for various diseases. But Pomberger sees potential applications in other areas, including working with two material manufacturers on polymer drug delivery development. ReactWise's automation also includes software that can interface with robotic lab equipment to enhance precision in drug manufacturing. While they don't make the robotic equipment themselves, they can drive it if clients have it available. Founded in July 2024, ReactWise has 12 pilot trials underway with pharmaceutical companies, expecting the first full-scale deployments of their subscription software later this year. They're working with some Big Pharma players, though names haven't been disclosed yet. ReactWise recently raised $3.4 million in a pre-seed round, including $500,000 from YC and a £1.2 million grant from Innovate U.K. The rest comes from unnamed venture capitalists and angel investors committed to advancing AI-driven, sustainable pharmaceutical manufacturing. While ReactWise focuses on a specific part of the drug development chain, Pomberger believes speeding up this process can significantly reduce the time it takes to bring new drugs to patients. A typical drug development timeline is 10 to 12 years, with process development taking one to two years. By reducing this by an average of 60%, ReactWise can make a real difference. Other startups are applying AI to different aspects of drug development, so there's potential for compounding effects as more automation innovations are integrated. But when it comes to drug manufacturing, Pomberger believes ReactWise is leading the way. They compete with legacy software like JMP and a few others using AI, but ReactWise's access to high-quality datasets on chemical reactions sets them apart. "We are the only ones generating these high-quality datasets in-house," Pomberger said. "Most competitors provide software and instructions based on inputs. But we offer pretrained models that fundamentally understand chemistry. Clients can tell us their reaction of interest, hit start, and we give them process recommendations from day one, based on all the pre-work we've done in our lab. That's something nobody else is doing right now."

Related article

Comments (49)