Google Disbands AlphaFold Team Amid Shift in Research Strategy

The award-winning team behind Google DeepMind’s groundbreaking protein structure prediction tool, AlphaFold, has been disbanded. The move follows a significant shift in the company’s research strategy, with most key members reassigned and several others leaving the organization altogether.

In recent months, nearly a quarter of the full-time researchers involved in the project have departed from Google DeepMind. This includes Nobel laureate John Jumper, who announced his departure for Anthropic in June, along with two AlphaFold colleagues who followed him to the rival research lab.

The reorganization is part of a broader shift at Google DeepMind, which has been redirecting its resources towards Gemini-focused projects. According to Pushmeet Kohli, vice president of research at Google DeepMind, ‘Our strategy over the last nine years has been to focus on grand challenges… The strategy has evolved.’

AlphaFold’s impact cannot be overstated. Developed by Google DeepMind in 2018 and recognized as a solution to the protein folding problem in 2020, the tool predicts the three-dimensional structure of proteins from their amino acid sequences in minutes rather than years. This capability has revolutionized fields such as drug discovery, vaccine development, and research into neurodegenerative diseases like Alzheimer’s and Parkinson’s.

The AlphaFold team took advantage of decades’ worth of accumulated work on protein structures to train its AI model. By leveraging this existing knowledge, the researchers were able to create a tool that could predict the structure of proteins with unprecedented accuracy. The resulting database has been made available for free to academic researchers worldwide, providing them with access to over 200 million predicted structures.

The Nobel Prize in Chemistry awarded to Demis Hassabis and John Jumper in 2024 marked a significant milestone for AlphaFold’s creators. However, the recognition also seems to have had an unexpected consequence: several key team members are now leaving Google DeepMind or being reassigned within the company.

According to one DeepMind employee, ‘The three departing researchers were core members whose exits caused surprise inside the company.’ They have landed at Anthropic, a rival research lab that has recently opened a workbench focused on biology and drug discovery. This development raises questions about what kind of projects are now being prioritized by Google DeepMind.

The shift towards Gemini-focused projects is part of a broader effort to automate parts of the scientific process itself. By building systems powered by Gemini, researchers aim to assist scientists broadly and eventually streamline their work. However, this new strategy has also led to concerns about the industry’s emphasis on short-term results over long-term research goals.

The reorganization at Google DeepMind is not without controversy. Some see it as a natural consequence of AlphaFold having achieved its goal: solving the protein folding problem to a useful standard and making the database public and permanent. Others, however, view it as a sign that the industry now rewards short-term gains over sustained research efforts.

The outcome has significant implications for researchers weighing their options in the field. With nearly a quarter of AlphaFold’s original authors leaving Google DeepMind, many are left wondering what kind of projects will be prioritized by the company going forward.