Hot topic: Betting big on AI-based drug discovery

It’s no secret that 2025 is turning into a tricky year for life science companies looking to raise cash. Political and economic uncertainty has fed through into  a ‘risk-off’ environment. 

So, when DeepMind spin-out Isomorphic Labs, a London-based AI drug development company, announced last week (March 31) that it had pocketed a huge $600 million in its first external funding round, the news certainly made headlines. 

There are good individual reasons why the company, headed by DeepMind founder Sir Demis Hassabis, was able to raise so much money. His reputation is one of them. 

Another is its backing by platinum-tier investors including Alphabet (the majority shareholder) Google Ventures and Thrive Capital – the last being one of the biggest backers of ChatGPT maker OpenAI. 

Isomorphic Labs’ IP, building on DeepMind’s Nobel Prize winning AlphaFold technology, which predicts the structure of proteins, is a third. 

However, it still isn’t a given that AI will prove the game-changing drug development force that it’s been touted to be. There are still plenty of dissenting voices who argue it’s been hyped up.  

They say its utility is limited by the sheer complexity and unpredictability of what might be called “real life biology”. They don’t tend to deny it will have a significant place in drug discovery and development, but say its advocates exaggerate what it’s really capable of. 

Currently, there are no drugs developed by generative AI that have received marketing authorisation from regulators such as the FDA or EMA.  

One of the closest to the finishing line is Insilico Medicine’s rentosertib, a small molecule inhibitor for the treatment of idiopathic pulmonary fibrosis (IPF), which has completed Phase IIa trials. 

So, that $600 million investment in Isomorphic Labs represents a very, very big bet – especially considering the challenging market backdrop. 

Advocates, meanwhile, say we are only beginning to develop really capable AI drug development technologies. They point out there has already been demonstrable success – for example, with Benevolent AI’s identification of the pre-existing drug baricitinib as a potentially useful treatment for Covid – and say critics do not appreciate just how fast things are moving. 

Regardless of one’s views on the true potential of AI for drug discovery, it is heartening to see investors are still willing to put their hands in their pockets to fund transformational “jam tomorrow” technologies, at a time when safer bets tend to be the order of the day.