BenevolentAI draws $90M from Temasek amid reports of halved valuation
Days after news leaked that BenevolentAI had secured funding at a rate that would halve its $2 billion valuation, the machine learning startup officially announced it raised $90 million from Singapore-based Temasek for a minority stake.
The news deals another blow to embattled fund manager Neil Woodford. BenevolentAI was one of two biotech AI startups, along with Oxford Nanopore, Woodford was reportedly prepping for auction to raise cash for his still frozen Equity Income Fund. Nanopore’s $1.5 billion valuation remains steady, affirmed by fellow investor IP Group.
For Woodford, the damage may have actually already been done. Last week, Woodford Patient Capital Trust announced they were forced to write down the value of one of its assets, costing the trust’s net value 4 pence per share and an estimated $52 million. The trust has now revealed that unnamed investment is BenevolentAI.
Nevertheless, BenevolentAI’s deflated value may cut into the price Woodford ultimately fetches for his stake in the company. An FT investigation from June found Woodford’s two firms collectively owned nearly a fifth of the Cambridge-based biotech. The biotech, in turn, makes up 5,04% of the Equity Income Fund and, per FT, is Woodford Patient Capital Trust’s largest single holding, accounting for 9.81%.
Woodford has said he plans to have enough cash to reopen the Equity Income Fund by December.
BenevolentAI’s value rose exponentially after Woodford’s initial investment in October, 2014. By its next funding round in 2015, it had reached unicorn status, rising from, £190m to £1.16bn, accounting for 2.86% – nearly one third – of the company’s 15.9% return that year, according to numbers cited by FT. By April 2019, less than two months before Woodford was all but forced to suspend his fund, it had reached $2 billion.
BenevolentAI and other machine learning startups have been the buzz of biotech the last few years, with their tantalizing promise of cutting into the steep, now on average over-$2.5 billion cost it takes to develop a new drug. Exact models vary by company, but broadly they promise to create more targeted R&D by scanning the vast troves of scientific literature and molecular databases no human or group of humans could manage and using it to find patterns and deduce which drugs or compounds are most likely to be effective. This month, a study from a smaller AI company, Insilico Medicine, made both industry-focused and mainstream headlines when it showed its technology could not only scan but also invent new molecules tailored to tackle a disease that proved effective in cell cultures and mice.
BenevolentAI appears to have similar software. It distinguishes itself by focusing on rare diseases and by the research facility the company acquired in Cambridge last year, which it says will equip staffers to develop the drugs from identification by AI systems to clinical research.
The company has partnerships with AstraZeneca and Novartis to develop drugs for chronic kidney disease and oncology. In a statement, the company said the cash infusion will allow it to “scale” its platform for discovery and development and those collaborations will bring in “meaningful revenue” over the next few years.
Social image: BenevolentAI via YouTube