avalw news
Liam ScottLiam ScottVIEW PROFILE →

DeepMind's Pivot: The Nobel-Winning Lab Behind AlphaFold Changes Course

tech2026-08-31 · 3 min read · 0 reads

Google DeepMind, the London lab that won a Nobel Prize for AlphaFold, has disbanded the team behind it in 2026 as Nobel laureate John Jumper departs for Anthropic. The lab is pivoting toward commercial general-purpose systems, while drug research continues at Isomorphic Labs.

Google DeepMind, the London based laboratory that redrew the boundaries of science, is changing direction. In 2026, the lab that won a Nobel Prize for cracking one of biology's hardest problems is steering away from pure research and toward the commercial technology its parent company is racing to defend.

The end of a landmark team

The clearest sign of the shift came from inside the lab itself. In 2026, DeepMind disbanded the team behind AlphaFold, the system that made its scientific reputation, marking a significant change in the company's underlying priorities.

Key figures headed for the door. The Nobel laureate John Jumper announced in June that he was leaving to join the rival lab Anthropic, and two of his AlphaFold colleagues soon followed him out the door.

Why AlphaFold mattered

AlphaFold predicted the structure of virtually every known protein, transforming structural biology.
AlphaFold predicted the structure of virtually every known protein, transforming structural biology.

To understand the significance, it helps to recall what AlphaFold achieved. In October 2024, Demis Hassabis and John Jumper won the Nobel Prize in Chemistry for the system, a rare honour for work rooted so deeply in computing.

The breakthrough reshaped an entire field. AlphaFold predicted the structure of virtually every known protein, solving a problem that had stumped biologists for decades and transforming structural biology almost overnight.

That success gave DeepMind a distinctive identity. It became known not just as a technology company but as a genuine scientific powerhouse, capable of discoveries that reached far beyond the usual boundaries of the software industry.

The pull of the general-purpose path

Now that identity is being reshaped. The lab has increasingly favoured what insiders describe as the general-purpose path, work that is broader, faster to deploy and tied directly to the revenue that its parent Alphabet is fighting hard to protect.

That choice has subordinated the scientific path. The pure research that built DeepMind's fame now sits beneath the commercial priorities that dominate the wider industry, a reordering that has unsettled some within the science community.

The lab has confirmed the change of course. DeepMind says its strategy has shifted, with researchers now channelled into projects tied to its Gemini models, alongside work on enzyme design, nuclear fusion, genomics and drug discovery.

Science lives on in a spin-off

The scientific mission has not vanished entirely, though. Much of the drug discovery work now runs through Isomorphic Labs, the biopharmaceutical spin-off that DeepMind created to turn its research into real medicines.

That venture is still pushing the frontier. Nearly two years after the release of AlphaFold3, which was geared toward drug discovery, Isomorphic Labs unveiled an even more powerful model, a drug-discovery engine it calls IsoDDE, in a detailed technical report published in February 2026.

A defining choice

The pivot reflects a deep tension at the heart of modern technology. Cutting edge science wins prestige and prizes, but the money and the competitive pressure increasingly lie in general-purpose systems that can be sold at enormous scale.

For DeepMind, the stakes are high on both sides. The lab is betting that it can serve its parent's commercial needs without losing the scientific spark that set it apart, a balance that will shape not just its own future but the direction of one of the world's most influential technology companies.

Liam Scott
Stay updated
Liam Scott
Subscribe to get an email whenever Liam Scott publishes a new story. No spam, unsubscribe anytime.
Liam Scott
WRITTEN BY THE AUTHOR
Liam Scott
2026-08-31 · 3 min read · 0 reads
View profile →
VERIFY THIS STORY
ASK AI
MORE FROM Liam Scott
Report this articlesupport@avalw.com