From OpenAI’s offices to a deal with Eli Lilly — how Chai Discovery became one of the flashiest names in AI drug development

Drug discovery, the artwork of figuring out new molecules to develop prescription drugs, is a notoriously time-consuming and troublesome course of. Conventional strategies, like high-throughput screening, provide an costly scattershot strategy — one that isn’t usually profitable. Nonetheless, a brand new breed of biotech corporations are leveraging AI and superior information applied sciences in an try and speed up and streamline the method.

Chai Discovery, an AI startup based in 2024, is one such firm. In a little bit over 12 months, its younger co-founders have managed to boost a whole lot of thousands and thousands of {dollars} and rally the backing of a few of Silicon Valley’s most influential buyers, making it one of many flashiest companies in a rising business. In December, the corporate accomplished its Collection B, bringing in an extra $130 million and a valuation of $1.3 billion.

Final Friday, Chai additionally introduced a partnership with Eli Lilly, a deal through which the pharmaceutical large will use the startup’s software program to assist develop new medicines. Chai’s algorithm, referred to as Chai-2, is designed to develop antibodies — the proteins essential to battle sicknesses. The startup has mentioned it hopes to function a form of “computer-aided design suite” for molecules.

It’s a important second for Chai’s specific area. The startup’s deal was introduced shortly earlier than Eli Lilly mentioned it will additionally collaborate with Nvidia on a $1 billion partnership to create an AI drug discovery lab in San Francisco. This “co-innovation lab,” because it’s being referred to as, will mix massive information, compute assets, and scientific experience, all in an try and speed up the velocity of latest medication improvement.

The business isn’t with out its detractors. Some business veterans appear to really feel that — given how troublesome conventional drug improvement is — these new applied sciences are unlikely to have a significant impression. Nonetheless, for each naysayer, there appear to be simply as many believers.

Elena Viboch, managing director at Basic Catalyst — one in every of Chai’s main backers — advised Trendster that her agency is assured that corporations that undertake the startup’s companies will see outcomes. “We imagine the biopharma corporations that transfer probably the most shortly to associate with corporations like Chai would be the first to get molecules into the clinic, and can make medicines that matter,” Viboch mentioned. “In apply which means partnering in 2026 and by the top of 2027 seeing first-in-class medicines enter into scientific trials.”

Aliza Apple, the pinnacle of Lilly’s TuneLab program — which makes use of AI and machine studying to advance drug discovery — additionally expressed confidence in Chai’s product. “By combining Chai’s generative design fashions with Lilly’s deep biologics experience and proprietary information, we intend to push the frontier of how AI can design higher molecules from the outset, with the final word objective to assist speed up the event of progressive medicines for sufferers,” she mentioned.

Chai might have been based lower than two years in the past, however the startup’s origins started round six years in the past, amid conversations between its co-founders and OpenAI CEO Sam Altman. A kind of founders, Josh Meier, beforehand labored for OpenAI in 2018 on its analysis and engineering workforce. After he left the corporate, Altman messaged Meier’s outdated faculty buddy, Jack Dent, to ask a couple of potential enterprise alternative. Meier and Dent had initially met in laptop science courses at Harvard however, on the time, Dent was a Stripe engineer (one other firm Altman was an early backer of). Altman requested him if he thought Meier can be open to collaborating on a proteomics startup — that’s, an organization centered on the research of proteins.

Altman “messaged me to say that everybody at OpenAI thought extremely of him and requested if I assumed he’d be open to working with them on a proteomics spinout,” Dent mentioned. Dent advised Altman “in fact,” however there was only one hitch: Meier didn’t really feel just like the know-how was fairly “there” but. The AI tech behind such companies — which leverage highly effective algorithms — was nonetheless a rising area and much from the place it wanted to be.

Meier was additionally fairly lifeless set on becoming a member of Fb’s analysis and engineering workforce, which is what he would go on to do. At Fb, Meier helped to develop ESM1, the primary transformer protein-language mannequin — an vital precursor to the work Chai is at the moment doing. After Meier’s time at Fb, he would spend three years at Absci, one other AI biotech agency based mostly round drug creation.

By 2024, Meier and Dent lastly felt ready to sort out the proteomics firm they’d initially mentioned with Altman. “Josh and I reached again out to Sam and advised him we should always decide up that dialog the place we left off — and that we had been beginning Chai collectively,” Dent mentioned.

OpenAI ended up turning into one in every of Chai’s first seed buyers. Meier and Dent really based Chai — together with their co-founders, Matthew McPartlon and Jacques Boitreaud — whereas figuring out of the AI large’s workplaces in San Francisco’s Mission neighborhood. “They had been form sufficient to provide us some workplace house,” Dent revealed.

Now, a little bit over a yr later, as Chai basks within the glow of its newfound partnership with Eli Lilly, Dent says that the important thing to the corporate’s quick development has been assembling a workforce of vastly proficient folks. “We actually simply put our heads down and pushed the frontier of what these fashions are able to,” mentioned Dent. “Each line of code in our codebase is homegrown. We’re not taking LLMs off the shelf which might be within the open supply [ecosystem] and fine-tuning them. These are extremely customized architectures.”

Basic Catalyst’s Viboch advised Trendster that she felt Chai was able to hit the bottom operating. “There are not any elementary obstacles to deployment of those fashions in drug discovery,” she mentioned. “Firms will nonetheless must take drug candidates via testing and scientific trials, however we imagine there’ll be vital benefits to those that undertake these applied sciences — not simply in compressing discovery timelines, but additionally in unlocking courses of medicines which have traditionally been troublesome to develop.”