There are few startup bosses with the kind of financial security Isomorphic Labs president Colin Murdoch enjoys.
The London-based AI techbio, which spun out of Google DeepMind in 2021 to commercialise a drug discovery platform, was backed for the first few years of its life by the very deep pockets of parent company Alphabet, which pumped hundreds of millions of pounds into the company.
In May, the company raised external funds for the first time, raking in $600m from the likes of OpenAI-backer Thrive Capital and Alphabet’s venture arm GV.
Isomorphic will need all that cash and a considerable amount more as it looks to slash the decade or so it takes to develop many new drugs by half — and cut the billions of dollars pharma companies spend on R&D.
It’s a big market opportunity, says Isomorphic Labs president Murdoch, but not one that generates quick returns.
“What we’re doing is a very long term thing,” says Murdoch. “We’re building a platform to develop medicines and if we can transform the drug discovery process even a little bit, it can have huge impacts for pharma companies.”
Spinning out of DeepMind
Isomorphic Labs is the brainchild of Murdoch and DeepMind founder and CEO — and recent Nobel prize winner — Demis Hassabis. The company is building a drug discovery platform using its sister company’s groundbreaking AlphaFold technology, an AI system which predicts the structure of proteins.
“There was this research breakthrough called AlphaFold and I remember thinking ‘What am I going to do with this? It’s incredible, but it doesn’t really fit,’” says Murdoch. DeepMind’s mission is to develop artificial general intelligence, but AlphaFold was a domain specific tool.
“After a few conversations with Demis, we got interested about something else we could do here,” Murdoch, who joined DeepMind as chief business officer in 2015, tells Sifted. It was the first time DeepMind had developed a technology with such immediate commercial potential, he adds.
Early discussions involved looking into applying AlphaFold commercially in sectors like agriculture and chemical synthesis, but it was drug discovery the company settled on.
The process of bringing a drug to market can be “tortuous” Murdoch says. It can take 10-15 years and cost billions of dollars to bring a drug to market. Even after all of the resources poured into R&D, just one in ten drugs make it through human trials.
Commercialising AlphaFold
The decision to spin Isomorphic Labs out of DeepMind was made for a few reasons, Murdoch says.
The company needed a different mix of staff than its sister company. DeepMind at its core was a mix of machine learning researchers, engineers and product managers, he adds. If you’re designing drugs, you need chemists, pharmacologists and biologists, says Murdoch, and there wasn’t even a career ladder for these types of people at DeepMind.
Spinning out allowed the company to set a clearer mission than it would have been able to in DeepMind, he adds, as well as allowing it to create a distinct culture from its sister.
“A dedicated mission and purposeful culture gives us the best chance of success,” he tells Sifted. “If you try to do too many things under the same umbrella you end up diluting.”
Murdoch, who still holds the role of chief business officer at DeepMind, spends nearly all of his time on Isomorphic Labs now. Hassabis, who is CEO at the techbio and DeepMind, works on the project about 10-20% of his time, Murdoch says.
Isomorphic Labs has already had some commercial success. At the start of 2024 the company inked two deals with pharma giant Novartis and Eli Lilly, which netted it $82.5m upfront and could be worth nearly $3bn further down the line.
The company follows in the wake of two ultimately unsuccessful British AI techbios, Benevolent AI and Exscientia, which both underwent ill-fated IPOs that saw share prices tumble, workforces slashed and, eventually, cut price mergers with other firms.
Advancements in AI since then are important, says Murdoch. “In many ways they blazed a trail. But the quality of machine learning research and maturity of AI tech now is on another level.”
That progress tempted external investors Thrive and GV. While Alphabet has bags of cash at its disposal, those two backers bring strategic benefits, says Murdoch.
GV has big biotech experience (the investor has backed 133 companies in the sector, according to Dealroom) which can help with strategy and connecting to potential commercial partners, he adds.
“Thrive has a great deal of experience growing next category companies, and we’re clearly not classic biotech or tech, we’re something new,” Murdoch says.
Talent hubs
One of the big challenges building an AI techbio is combining top AI workers with pharma veterans and scientists, says Murdoch. “We’re blending two worlds that haven’t really come together.”
Access to both pools of talent is why Isomorphic Labs — which is building its own foundational models — is headquartered in London, he tells Sifted.
London has the world’s best AI talent — driven by a combination of world-leading universities as well as DeepMind’s talent flywheel, Murdoch says. The city is also a leader in the life sciences space, he adds.
It means that while Europe has seen a flight to the US, in particular to San Francisco, from some of its most promising AI startups in recent times, Murdoch’s not considering upping sticks — though the company did open a small office in Boston, Massachusetts, earlier this year.
SF might have AI talent on a par with London, but it lacks biotech expertise on a level with the British capital Isomorphic Labs needs.
“It’s the perfect location for building a company like this, because you’ve got the combination of AI and life sciences all under one roof,” says Murdoch. “You have these centres of biotech and tech around the world but often they’re not in the same place.” Isomorphic Labs also has a presence in Lausanne, Switzerland to access the former.
The future
Isomorphic Labs’ $600m fundraise will broadly be spent on two areas, developing its drug design engine (the AI system that develops new medicines) and expanding the number of drug design programmes it’s using its tech on.
Alongside projects with Novartis and Eli Lilly, Isomorphic Labs is working on its own drug development programmes in the cancer space. The company is “constantly” in discussion with other big pharma companies, says Murdoch, and is in talks to expand its partnership with Novartis.
There will also be further external fundraises down the line. While Isomorphic Labs isn’t actively raising at the moment, Murdoch says, he gets calls from investors every week. “I always take them because I want to build a relationship with them. It’s important to stay in touch and hear what they’re interested in.”
The company’s headcount is now approaching 250 and it plans to grow by another 100 or so in the next 12 months as it looks to hit the next big milestone towards getting an AlphaFold-designed medicines on the shelves — a drug candidate entering clinical trials
While Murdoch says that stage is “very close” — pointing to the company hiring a chief medical officer in the US in preparation — he’s coy when asked to put a more specific timeframe on it.
Earlier this year, Hassabis said he expected clinical trials by the end of 2025. Whether that target is hit remains to be seen, but advances in AI are undoubtedly ushering in a new age of techbio.
“AI technology is now able to do what used to take six months in a wet lab six minutes on a computer,” Murdoch says.
Read the orginal article: https://sifted.eu/articles/isomorphic-labs-big-interview/
