Plotting to be the BridgeBio of AI, Atomwise lands $123 million Series B for hype-heavy platform
The PR-friendly, well-partnered AI biotech that’s provoked stern skepticism in some scientific corners is getting a boatload of new cash.
Atomwise has announced a $123 million Series B round led by Sanabil Investments — a subsidiary of the Saudi royal fund — and B Capital Group and joined by DCVC and Y Combinator, among others. The new round is nearly triple what Atomwise had raised prior and will go towards both scaling their molecule-hunting software and building the growing network of spinouts they’re launching to develop some of the molecules that software has turned up.
The goal ultimately, said CEO Abe Heifets, is to build a portfolio of smaller biotechs beneath theirs — a kind of BridgeBio for AI.
“We want to grow in scale,” Heifets told Endpoints News. “The technology is small molecule — that’s a very broad umbrella so there’s increasingly an interest in a portfolio approach”
The round could be a legitimating one for Atomwise, a company that over the last few years has found itself at the center of a debate between engineers who promised that machine learning and AI networks could remake drug development and scientists who saw a lot of buzz but little substance. Since its days at Y Combinator, the company has promised to use an AI convolutional network to rapidly screen billions of molecules for their ability to hit a target or bind to a protein, and in doing so speed from “years to days” the process of selecting drug candidates. In doing so, they said, they could cut short the arduous and expensive drug development path.
The problem, critics such as science blogger and medicinal chemist Derek Lowe argued, is that it just doesn’t take that long to screen molecules. It’s a bump in the drug discovery mountain.
“You can do a million in six weeks. The whole compound screening step is just another early thing in preclinical space,” Lowe wrote in one piece that also noted Atomwise’s “tendency toward overstatement.”
“I’ve never seen a successful project in which it was a rate-limiting step. But ‘shave a few weeks off something at the very beginning’ isn’t as compelling an offer, is it?” he said.
Though just one of many companies now offering rapid, AI-enabled screening, Atomwise might be the most prolific, claiming “over 750 research collaborations addressing over 600 disease targets” and partnerships with major pharma companies, including Eli Lilly, Merck, Korea’s Hansoh Pharmaceuticals, Bayer and BridgeBio.
Yet it has advertised those big-name partnerships with particular fanfare. Rather than calculate the overall potential deal value by upfront fees and milestones, as most biotechs do, they have often listed values that include royalty estimates “based on historical average revenues for small molecule drugs … with success in all projects” — in an industry where success is fleetingly rare. That’s allowed the company to advertise that “Atomwise has signed more than $5.5 billion in total deal value with corporate partners to date” without disclosing any individual payments.
Heifets says that their deal releases are in line with how other biotechs talk about their deals and what their partners are willing to disclose. He also says Atomwise provides benefits beyond that initial screening step.
“If you think of AI as only being applied for high throughput screening, then I agree with Derek,” he said. “That’s a beginning part and that’s pretty quick.”
Heifets said that Atomwise also provides services for lead optimization, a longer and more difficult step. And he said that they’ve shown the software can not only find molecules faster but also find molecules for targets that major companies have spent years and millions of dollars failing to hit.
Most notably, Atomwise launched X-37 last year in part around the discovery of molecules that can bind to the PIM3 pathway without harming healthy tissue, which Heifets said Roche, Novartis and AstraZeneca had tried and failed to do. The company raised $14.5 million in Series A funding. They also quietly launched Theia Biosciences around molecules that can hit the HTRA1 pathway and might be used to treat age-related macular degeneration.
Over the next few years, Heifets said, they plan to launch more biotech subsidiaries, hopefully eventually putting multiple drugs in the clinic. If the markets keep their historic pace, he said, an IPO could also be in their fortunes.
That would likely mean a hefty S-1 and another round of arguments over the role of AI in biotech, and what’s hype and what’s reality.
“I think that will depend on what the markets are doing,” Heifets said. “There have been a number of very successful IPOs recently in the biotech sector, so it’s a very interesting time.”