Alphabet's Isomorphic Labs in Talks to Raise About $5 Billion for AI Drug Discovery

The round has not been finalised and its terms could still change, according to the report. Coverage that followed Bloomberg's story put the potential valuation at about $50 billion. No investors in the new round have been named.

If it closes at that size, the round would be one of the largest private raises by an AI company outside the frontier model labs, and the largest yet for a company using AI to design medicines. It would also be more than twice the $2.1 billion Isomorphic raised earlier this year. Reports differ on when that round was announced, with some putting it in March and others in May. It included Alphabet, two of Alphabet's venture funds and outside backers, and has been reported as led by Thrive Capital.

The talks appear to have moved on quickly. On October 8, Bloomberg reported that Isomorphic was in early discussions about funding at a valuation of between $40 billion and $50 billion, but that report gave no size for the round.

Isomorphic is led by Demis Hassabis, who also runs Google DeepMind and shared the 2024 Nobel Prize in chemistry for the AlphaFold protein structure prediction work on which the company's science is built. Its in-house platform, called IsoDDE, is designed to automate early drug discovery steps such as finding binding pockets, the sites on a protein where a drug molecule can attach. These are hard to pin down because they can shift or disappear as the protein changes shape.

In February, the company said IsoDDE was more than 50 per cent better at finding binding pockets than the closest competing tool. It also said the platform predicted how drugs bind to proteins about twice as accurately as AlphaFold, and modelled antibody interactions about 2.3 times better. Those figures are the company's own and have not been independently verified.

Isomorphic makes its money mainly from partnerships with drug makers. In 2024 it signed deals with Eli Lilly and Novartis worth nearly $3 billion in total, counting potential milestone payments, and in January it began a collaboration with Johnson & Johnson. All three deals cover small-molecule drugs.

The size of the round would give Isomorphic the money to take more of its own drug candidates towards clinical testing, which costs far more than computational discovery work. It would also make it easier to buy the computing capacity its models need, which has become expensive as AI labs compete for chips. The company has not said publicly when its first internally discovered drugs will enter human trials.

Isomorphic and Alphabet had not commented publicly on the reported talks.