SoftBank, the Japanese telecommunications giant, has been on a multibillion-dollar acquisition and investment spree for the past year or so. It isn’t over. Today the pharmaceutical company Roivant is announcing a $1.1 billion investment, mostly from the SoftBank Vision Fund. It’s perhaps the largest single biotech investment ever. And after the biggest biotech IPOs of 2015 and 2016, this new money might also be another validation of Roivant’s development approach, grabbing up promising-but-abandoned drug candidates that other companies couldn’t get across the finish line.
A quick tally: SoftBank bought the robot maker Boston Dynamics from Google parent company Alphabet. It bought the chipmaker ARM for $32 billion. Through its $100-billion investment fund (with contributions from Apple and Oracle, among others) SoftBank put $4 billion into the chipmaker Nvidia, $130 million into the gene-modifying biotech Zymergen, $1 billion into sports paraphernalia company Fanatics, and at least $300 million in co-working developer WeWork. It is a portfolio diverse to the point of weirdness.
And now this firehose of money is spraying onto a new world. Most of the $1.1 billion is coming from SoftBank; other than confirming the investment, a representative from SoftBank declined to identify who was fronting the remainder. A Roivant spokesperson said the bulk of the rest was coming from existing Roivant investors.
Just three years old, Roivant primarily works through a collection of subsidiaries developing different drugs, all originating from carefully negotiated raids of other pharma companies’ cupboards. Myovant is working on a drug for endometriosis and uterine fibroid pain. Urovant is going after incontinence and overactive bladder. Axovant is chasing dementia and Alzheimer’s. And so on—drugs rescued from abandonment and then put through fast-moving tests to get them commercialized.
Several of the drugs are in Phase III trials, which means they’re being tested on humans for efficacy, one of the final steps before government approval. Several more are in Phase II. “It’s the pipeline of a large pharmaceutical company at about a fiftieth of the burn that any of those companies would spend to get to that same point,” says Barry Selick, a longtime biotech executive now vice chancellor for business development, innovation, and partnerships at UC San Francisco. “It’s pretty remarkable.”
Roivant’s CEO, Vivek Ramaswamy, studied molecular biology as an undergraduate but then headed for the finance business, focusing on biotech. That’s where he started to think the process of turning interesting chemicals into drugs didn’t really work—it took decades and cost billions, and even after all that a drug might still fail out. Worse, the ways other industries have learned to use data to optimize their processes were still largely foreign.
Ramaswamy suspected he could cut through that by starting where other companies had decided to quit. Roivant goes looking for drugs stuck in turnaround—not because of problems with the science, but because of corporate changes-of-plan. “Once we take over those drugs, the same cultural attributes that allowed us to focus on specific drugs rather than general therapeutic categories allow us to focus on the process by which those drugs can be accelerated to the finish line,” Ramaswamy says.
The hard part, you’d think, would be knocking on some company’s door and saying, hi, can we have your stuff? But Ramaswamy says companies with drugs languishing on shelves are open to the idea. “They’re not spending any more money on it. They don’t have to dilute corporate focus. And they can participate in the upside with us in the form of royalties or other arrangements,” he says. “Many of them are motivated by the idea that the work they put in has the potential to help patients. It’d be a shame for it to go to waste.”
To find the drugs, Roivant has relied on the kind of intelligence-gathering you might expect—public databases and networks of relationships. But there’s even more data out there. Using a strategy they call “mapping the drugome,” Roivant’s business development team digs into that data. “You can start seeing patents. You can identify companies where they were running clinical trials in a therapeutic area at a pretty steady state, and now all of a sudden that’s gone down to five, then three. That’s a trend,” says Dan Rothman, Roivant’s chief information officer.
The company’s analysts further chart those drugs against existing categories of drugs, methods of action, and endpoints that others in the business are pursuing. Roivant also looks at whether a given drug might have a patient population that needs it. That’s commercial viability. Rothman says Roivant has mapped 30,000 potential drugs, 2,000 mechanisms of action, and 10,000 endpoints this way—all from publicly available, mostly free databases.
In fact, that’s where some of the SoftBank money will go: to purchasing more expensive, presumably more valuable data on who’s developing what.
Roivant will also spend some of the new money on spinning up another subsidiary, Datavant, designed to improve the overall process of running clinical trials through data analysis. See, companies run a lot of trials and collect a lot of data, but they don’t publish and share it all. So researchers sometimes unknowingly replicate ideas that have already failed. “If all data was pooled, you could cut the cost of clinical trials roughly in half, you could reduce time to market by 50 percent, and significantly increase the odds of success of any given clinical trial,” says Travis May, president of the connectivity division of Acxiom and a member of Datavant’s board of directors.
It seems a little crazy, perhaps—get companies to share their darknets with each other? But Datavant isn’t the only place trying to do it. The US Department of Energy, National Cancer Institute, and pharma giant GlaxoSmithKline are trying, too, as part of the Accelerating Therapeutics for Opportunities in Medicine project. “GSK is looking for other pharmaceutical companies to share their data,” Selick says. Lawrence Livermore National Lab has the supercomputers that can then, perhaps, build models that can say what molecules, antibodies, and so on might be worth testing—useful not just for GSK but for everyone. “And the companies that are nimble, they’ll turn around and apply those predictive models to their own internal pipeline and programs.”
It’ll take time and the right people to build those models, of course. At Roivant, that’s why this latest round of investment is so welcome: more ‘vants, more engineers, more data. For its part, SoftBank has been largely mum on what that company gets out of disruptively investing in yet another sector—“we are impressed with the ambition and track record of the Roivant team and look forward to supporting them in the next step of their journey,” SoftBank Group International managing director Akshay Naheta said in a press release.
Here’s a clue, though: The company’s CEO Masayoshi Son sometimes says he’s preparing for the arrival of digital superintelligence. Well, if the singularity speeds up the creation of new drugs people need, bring on the robot overlords.