Exscientia’s latest €434.9M ($525M) financing last month showed investors are still excited by efforts to use artificial intelligence to fuel drug discovery, with global tech investors often taking the lead.
The Oxford, UK-based company Exscientia has one of the most mature artificial intelligence (AI)-focused drug discovery pipelines, alongside companies like BenevolentAI in the UK, and Aria Pharmaceuticals and Atomwise in the US.
Earlier this week, Exscientia announced its third partnered program was heading to the clinic, a small molecule for psychotic symptoms resulting from Alzheimer’s disease. This is the second clinical asset from Exscientia’s neuroscience partnership with the Japanese firm Sumitomo Dainippon Pharma — the first went to clinical trials last year for the treatment of obsessive-compulsive disorder. Exscientia’s other clinical asset, a cancer immunotherapy developed with the German company Evotec, began phase I testing in April.
Exscientia’s latest financing included a €186.4M ($225M) Series D led by the tech-focused SoftBank, which also pledged an additional €248.5M ($300M) equity commitment for use at the company’s discretion. Combined, that makes this the largest single venture raise for any company in the space and likely sets the company up for an initial public offering (IPO).
Ben Taylor, Exscientia’s CFO, says that alongside new collaborations like the Evotec and Sumitomo Dainippon Pharma ones, the funding will allow the company to rapidly scale its platform and pipeline “toward full end-to-end autonomous drug design.”
Tech VC investors have been branching out into the biotech sphere for years to back companies tapping AI, an umbrella term generally denoting self-adaptive machine learning algorithms for data analysis.
Drug discovery in particular has been a big draw, with startups eager to show they can adopt techniques from other sectors to accelerate, improve, and lower costs over traditional approaches. This has especially been the case during the pandemic, where AI-driven drug discovery companies have had the chance to prove the worth of their strategies against Covid-19, which has limited treatment options.
At least 42 companies worldwide are tapping AI for drug discovery, but of the ten biggest financing rounds for AI-drug discovery companies since 2018, five have had a tech-focused lead or co-lead investor, while only four had a biotech-focused investor leading the round.
The top rounds are also dominated by US fundraisers, with only Exscientia and BenevolentAI in Europe.
While several companies have AI-discovered assets in human testing, BenevolentAI is unique in having clinical validation for its AI-guided drug discovery approach.
BenevolentAI’s achievement stems from its prediction in February 2020 that Eli Lilly & Co’s approved rheumatoid arthritis drug Olumiant (baricitinib) would be effective in treating Covid-19. The US FDA granted Lilly an Emergency Use Authorization for the therapy in November following promising phase III data in hospitalized Covid-19 patients. Although the drug failed to meet the trial’s primary endpoint — reducing the risk of the disease worsening — there was a significant reduction in mortality from any cause.
Repurposing old drugs or chasing validated targets is just the first step for these companies. David Grainger, a co-founder and Chief Scientific Advisor at Medicxi, says AI is already well established in small molecule discovery for known targets, where it has been making substantial contributions for 20 years. The big bet comes in developing AI approaches for identifying new drug targets.
“That’s the only question in modern drug discovery that isn’t commoditized,” says Grainger.
It’s tricky, Grainger adds, because confirmation of a novel target hypothesis via clinical trials can take a decade – something lost in the hype from tech investors. AI approaches may indeed continue to improve discovery, but Medicxi has not been convinced by any company with AI discovery platforms as a core competency.
“Exscientia will have to spend the $525M to find out if they were right,” says Grainger. “My guess is, they will be incrementally better – not a revolution.”
Some traditional biotech backers have not had the same hesitation. For example, Blackrock joined Exscientia’s €82.8M ($100M) Series C round in March this year, as well as a €101.9M ($123M) Series B for Cellarity (a Flagship Pioneering startup) in February. And in May last year, OrbiMed co-led a €87.0M ($105M) Series B for Canada-based AbCellera, months before the company raised €458.7M ($555.5M) in an IPO in December.
While conventional biotech investors know what they’re getting into, Grainger thinks tech investors have looked at the space and come away with the mistaken impression that AI approaches successful in other industries can be plugged into drug discovery. “I’m a critic of the unthinking hype,” he says.
Exscientia’s Taylor says new approaches are needed to improve on biopharma’s 95% failure rate, and tech investors are not jumping in blind. “Some of the most in-depth and insightful biotech questions [we’ve gotten] actually came out of firms who were traditionally considered tech investors,” he says.
In addition, although the convergence of technologies has expanded the understanding among biotech investors in recent years, Exscientia was looking for a balance of specialists and long-term backers. “It was important to bring in investors who not only shared our vision, but also had the ability to support us for many years while we execute on our plan,” says Taylor.
Softbank, which led Exscientia’s mega-round, is one of the most prominent tech investors to delve into AI drug discovery, alongside Atomico, B Capital Group, The Chartered Group, Bulba Ventures, Drive Capital, and Koch Disruptive Technologies. It led or participated in other major rounds in the space including XtalPi, insitro, and Relay Therapeutics – all based in the US. Nonetheless, SoftBank’s recent focus on Exscientia could bode well for other European players – few of which have announced venture rounds to date.
Cover image from Elena Resko.