Some biotechs struggle for cash. Recursion, lately, seems to be swimming in it.
Today, having already raised over $180 million in the last two years, the Salt Lake City-based AI drug developer announced a $239 million Series D. That’s more than any US or European biotech has raised in a round this year and more than all but two biotechs raised last year. (Mabwell, a Shanghai-based antibody developer, raised $278 million in a Series A in April.)
The round came with an important point of validation for Recursion and a potential long-term source of revenue: a new partnership with Bayer. That deal, centered on fibrotic diseases, will pay Recursion $30 million upfront for the partnership, along with $100 million in milestones for each of up to 10 programs the companies could pursue. So it’s an up to $1 billion deal, even if it’s vanishingly unlikely to reach that. Leaps by Bayer, the German pharma’s venture arm, also led the Series D, contributing a $50 million equity investment.
The investment-partnership represents one of the largest rounds, if not the single largest round, for an artificial-intelligence focused biotech and cements Recursion as one of the major players in a nascent field that has produced many small startups but few heavyweights. It also points to where Big Pharma and major VCs may spread their focus as machine learning approaches advance. The first small companies to make headlines and sign deals used machine learning to screen vast libraries of molecules in search of ones that can hit targets that drug companies have long tried to hit. More recently, though, a couple upstarts have raised significant cash and scored prominent partnerships using advanced computational tools to study cells more closely and come up with new targets themselves.
“What Recursion does is really hard so I wouldn’t say there is going to be a deluge of companies,” Zachary Bogue, founder of early Recursion backer DCVC, told Endpoints News. “But this idea of biology as a platform and using AI as a drug discovery is the new frontier in biotech.”
Roughly, that’s what Recursion does. In a 100,000 square-foot warehouse in downtown Salt Lake City, robots take petri dishes of different cell types and knock out different genes, taking constant pictures in the process. Humans can’t easily tell the difference between most of those pictures, but computers can, and with enough images and hundreds of different measurements on each, they can pick up patterns to indicate what can make a cell sick and which genes, when targeted, can make them healthy. They can then identify and tweak molecules or compounds that hit those targets.
So far, they’ve used that approach to identify molecules to bring into clinical development for several rare neurological conditions and hereditary cancer syndrome, pulling compounds from Ohio State, Takeda, and co-founder Dean Li’s labs. But the company lists a bevy of preclinical disease areas on its pipeline, and Recursion CEO Chris Gibson said that, with the Bayer deal, they would begin to look to partner with big companies where clinical development is more complex, such as in neurology and oncology.
“We’ve been talking about this internally as the beginning of a second act for the company,” he said in an interview with Endpoints.
Other companies have scored large amounts of capital with similar approaches. Most notably, Daphne Koller’s startup Insitro has nabbed two $100 million-plus rounds in a span of 13 months, plus a Gilead collaboration on NASH. In January, star Canadian researcher Brendan Frey’s Deep Genomics raised $40 million into the clinic. Meanwhile, the rhetoric from the handful of drug pharma executives who talk openly about machine learning, such as GlaxoSmithKline’s Hal Barron, has centered on approaches that help uncover not just new molecules but also new targets.
Under the deal with Bayer, Recursion will use their system to build models for different fibrotic diseases, following guidance from Bayer fibrosis experts at the pharma. They will then use Bayer’s library of small molecules and their own internal ability to screen and develop molecules to come up with preclinical candidates. Around that point, Gibson said, they’ll hand things off to Germany.
Although Leaps by Bayer described their $50 million contribution to the Series D as “our big bet in terms of digital drug discovery,” Bayer itself has also invested in the molecule-screening side of the AI biotech world, as have other major companies. In January, Bayer and the UK-based Exscientia signed a discovery collaboration for up to €240 million. Exscientia has also teamed with Sumitomo, Sanofi and Celgene, the last of which included a $25 million upfront payment. Atomwise and Insilico have also signed multiple big-name partnerships, although they have been largely milestone-heavy, with little upfront disclosed.