ALPHAFOLD. NOBEL.
THE DEAL
Protein structure prediction had resisted researchers for decades.
SOURCES: Nobel PrizeNobel Prize
Jumper took the protein-folding problem back to grad school and came out with AlphaFold2 — predictions accurate enough to win CASP, the blind protein-folding contest, and end the game: a database covering almost every known protein, and a Nobel Prize. In 2026 he walked out of DeepMind into Anthropic, right in the middle of the frontier talent war.
2008
After studying physics and mathematics at Vanderbilt and theoretical condensed-matter physics at Cambridge, Jumper joined D. E. Shaw Research and worked on molecular-dynamics simulations of proteins.
SOURCES: Vanderbilt Universityreported
2011
Jumper entered the University of Chicago's doctoral program and began combining machine learning with protein folding and dynamics, the bridge into the work that later became AlphaFold.
SOURCES: Vanderbilt Universityreported
Dec 2, 2018
DeepMind's first AlphaFold system placed first at CASP13, establishing the project as a serious new approach to protein structure prediction.
SOURCES: Google DeepMindfirst-party
Nov 30, 2020
At CASP14, Jumper and the AlphaFold team produced predictions accurate enough for organizers to recognize AlphaFold2 as a solution to the protein-structure prediction problem.
SOURCES: Google DeepMindfirst-partyNobel Prizereported
Jul 15, 2021
The AlphaFold2 methods were published in Nature and DeepMind released code, followed days later by the AlphaFold Protein Structure Database.
SOURCES: Google DeepMindfirst-party
Jul 28, 2022
DeepMind expanded AlphaFold predictions to more than 200 million protein structures, covering nearly all catalogued proteins known to science.
SOURCES: Google DeepMindfirst-party
Oct 9, 2024
Jumper and Demis Hassabis shared half of the 2024 Nobel Prize in Chemistry for protein structure prediction through AlphaFold.
SOURCES: Nobel PrizereportedGoogle DeepMindfirst-party
Jun 19, 2026
Jumper announced he would leave Google DeepMind and join Anthropic, moving one of AI-for-science's biggest names into the frontier-model talent war.
SOURCES: Reutersreported
IDEA
A protein begins as a chain of amino acids but works as a three-dimensional object. AlphaFold2 attacks the hard mapping between those two descriptions, predicting structure from sequence with accuracy that transformed protein-structure prediction.
SOURCES: Nobel Prize
IDEA
The Nobel committee's account emphasizes that Jumper's protein knowledge helped reshape AlphaFold after the earlier system stalled. AlphaFold2 combined machine learning with biological and chemical insight rather than treating the problem as generic pattern recognition.
SOURCES: Nobel Prize
IDEA
The AlphaFold project did not stop at a benchmark win. DeepMind released the methods and code and, with EMBL-EBI, expanded predicted structures to more than 200 million proteins, turning one research result into a tool used across biology.
SOURCES: Google DeepMindNobel Prize
RECORD
In 2020 AlphaFold2 produced protein-structure predictions accurate enough that CASP organizers treated the long-running prediction challenge as effectively solved for many targets.
SOURCES: Google DeepMindNobel Prize
RECORD
DeepMind expanded AlphaFold predictions to cover virtually all catalogued proteins, giving researchers a structure resource at a scale that experimental methods could not approach.
SOURCES: Google DeepMind
RECORD
Jumper and Demis Hassabis jointly received half of the 2024 Nobel Prize in Chemistry for protein structure prediction through AlphaFold.
SOURCES: Nobel Prize
NOW
In June 2026 Jumper said he was leaving Google DeepMind after nearly nine years to join Anthropic. The move puts one of AI-for-science's most prominent researchers inside a frontier-model company that is simultaneously expanding its life-sciences ambitions.
SOURCES: Reuters
