Google Shuts Down Nobel Prize-Winning AI "AlphaFold" to Focus on Gemini
Google DeepMind has ended the AlphaFold project, shifting most team members to Gemini-related projects, while some have left the company.
Google DeepMind’s Nobel Prize-winning AI project for protein structure prediction, “AlphaFold,” has come to an end. According to a report by the Financial Times, Google has reassigned the majority of the project’s key members and original authors to other roles, while some researchers have already left the company. This move is seen as part of Google’s strategy to concentrate resources on its Gen AI initiative, “Gemini.”
AlphaFold is an AI program capable of predicting the 3D structure of proteins from amino acid sequences with remarkable accuracy, reducing what traditionally took years to mere minutes. It has been used to accelerate drug discovery, vaccine development, and the understanding of protein structure changes related to neurodegenerative diseases such as Alzheimer’s and Parkinson’s.
Solving a 50-Year Challenge
DeepMind began developing AlphaFold in 2018. By 2020, it was recognized as a groundbreaking solution to the “protein folding problem,” a challenge that humanity had grappled with for over 50 years. Protein 3D structures play a crucial role in determining their biological functions.
Over the past five decades, scientists have identified the structures of around 170,000 proteins using techniques like X-ray crystallography and nuclear magnetic resonance. The AlphaFold team leveraged this accumulated knowledge to train its AI models.
In 2021, the journal Nature published papers detailing AlphaFold’s methodology and its structural predictions for the entire human proteome. DeepMind also launched the “AlphaFold Protein Structure Database,” offering over 200 million protein structure predictions freely to researchers worldwide.
Challenges in Passing on Knowledge Post-Nobel
Recognition In 2024, DeepMind’s CEO Demis Hassabis and John Jumper, a researcher who joined at the project’s inception and later became VP and Engineering Fellow, were awarded the Nobel Prize in Chemistry for their contributions to AlphaFold.
However, Jumper announced in June 2026 that he would leave DeepMind to join AI company Anthropic, along with several colleagues. DeepMind confirmed that other team members have been reassigned to Gemini-related projects, while some were moved to Isomorphic Labs, a drug discovery company under Alphabet.
Pushmeet Kohli, DeepMind’s VP of Research, remarked, “Our strategy over the past nine years has been to focus on large-scale challenges, with every project having clear goals. However, our strategy is evolving.”
Accelerating Resource Allocation to Gemini
The Financial Times noted that the termination of AlphaFold symbolizes Google’s decision to focus resources on developing Gemini. Buoyed by the strong performance of its cloud business, the company has intensified its AI investments, prioritizing its foundational models.
While the open AlphaFold database remains accessible to researchers globally, the discontinuation of the project’s development and maintenance raises concerns about future updates and addressing new challenges.
Editorial Opinion
In the short term, the lack of clarity regarding the maintenance of the AlphaFold database presents a significant issue. Although the existing data remains valuable, further improvements in protein structure predictions or advancements in solving the protein folding problem may stagnate. The research community will likely look to open-source implementations or successor projects led by academic institutions for progress.
In the long term, while focusing resources on Gemini aligns with Google’s AI strategy, this move will be remembered as a case that questions the balance between contributions to fundamental science and business goals. AI applications in life sciences represent a unique domain that generates value beyond large-scale language models, and Google risks losing competitiveness in this field. Although research may continue under spin-offs like Anthropic or Isomorphic Labs, increased corporate competition may undermine the spirit of open science.
As an editorial stance, we believe that the dissolution of a project that produced Nobel Prize-worthy achievements due to a strategic pivot necessitates broader discussion and reflection.
References
- “Google shuts down its Nobel-prize winning AlphaFold project as it focuses on Gemini”, by staff@engadget.com (Mariella Moon) — Engadget, 2026-07-29T10:40:36.000Z (ARR)
- Source URL: https://www.engadget.com/2225849/google-shuts-down-alphafold/
Frequently Asked Questions
- Why was AlphaFold shut down?
- Google reallocated resources to focus on Gemini as part of its strategy, reassigned AlphaFold team members to Gemini-related projects, and some members transitioned to AI company Anthropic or drug discovery company Isomorphic Labs.
- Will the AlphaFold database remain accessible?
- The AlphaFold Protein Structure Database is still available for researchers. However, since the project’s development has ceased, new protein structure predictions or accuracy improvements may no longer be added.
- How will this impact life sciences research?
- In the short term, existing databases will continue to be utilized, but in the long term, concerns linger over the stagnation of open prediction models. The research community may look to successors from academic or corporate entities for innovation.
Comments