Google Shakes Up Leadership as Demis Hassabis Steps Back from DeepMind to All-In on AI Drug Discovery with Isomorphic Labs

The Internet Is Ablaze: Google Faces Another Personnel Earthquake

Previously,Demis HassabisDirect ResignationGoogle DeepMindChief Executive Officer,Stepping down from day-to-day management, transitioning to the role of Director and Chief Scientist, with core focus fully shifting towards AGI technology research and development and frontier scientific exploration.

Demis Hassabis stated,# TranslateContinue to serve asAIPharmaceutical CompanyIsomorphic LabsChief Executive Officer,will be invested in this company"A More 'Proactive' Role."

This seemingly sudden transfer of power is, in fact, a key signal of the iteration in Google’s AI strategy.

DeepMind bids farewell to the era of singular scientific exploration, fully embracing commercialization and the race for general AI, while Demis returns to his strongest suit: AI for Science.

As early as in public interviews, Demis had boldly predicted:

Over the next decade, leveraging disruptive breakthroughs of AI in the field of biology, humanity is expected to conquer major and intractable diseases such as cancer.

This move was actually foreshadowed long ago

This series of personnel changes was unexpected yet understandable.

Over the past month,Google DeepMindRepeatedly experienced personnel adjustments.

Forefoot,John JumperInXShangguan Xuan: Leaving work recently9of the yearGoogle DeepMind, joinAnthropic。

As a core key member working alongside Demis to tackle critical challenges, John Jumper led the AlphaFold project, which successfully solved the problem that had plagued biologists worldwide for 50 years.The Protein Folding Problem, fundamentally reshaping the paradigm of structural biology research, and the two were awarded the Nobel Prize in Chemistry for their groundbreaking contributions.

Last week,GoogleDeepMind# Exposed AgainHas built its portfolioAIProtein StructureAlphaFoldThe Behind-the-Scenes Team of the Series Disbanded.

The main reason lies in the fact that,GoogleDeepmindofAI4SThe team will no longer focus on a single scientific question such as protein folding, but instead will be dedicated to building aGeminiComplete Drive SetAIResearch System.

Meanwhile, Google beganActively Enter the MarketAIAgent, come withOpenAIandAnthropicCompetition.

In fact,Demis HassabisStepping down as CEO to focus on AI for Science, a move more within his comfort zone.

DeepMindHas been focused on since its inceptionAIScientific Questions: A Series of Milestone Models Including AlphaGo and AlphaFold

Unfortunately, these models cannot be monetized, and during difficult times, one can only5Sold to Google for a valuation of around hundreds of millions of dollars.

Fortunately, Google kept its promise after the acquisitionDeepMindGained Google's powerfulComputational resources and financial support,DeepMind retains its independent R&D framework, continuing to delve deeply into frontier exploration.

AndOpenAIThe emergence of which has made Google feel a deep sense of crisis, in 2023 willDeepMindWith Google Brain (Google Brain）Combined,Prioritize ensuring the Gemini model and its commercial implementation.

A number of Google researchers have stated that,Demis HassabisOn GoogleDeepMindMore efforts are devoted to the scientific team,Geminisuch specific products remainJeff DeanResponsible.

Therefore,Demis HassabisRemoveCEOthe burden, turning toIsomorphic Labs, it may be a better choice for him personally.

What’s more intriguing is that Google’s technological cornerstone, having joined the company 27 years ago,Jeff DeanAnnouncementinstead, he founded a startup named DiscoveryLoop.

The company will focus on researching methods to leverage artificial intelligence technologies for the automation of scientific and engineering tasks, which may include applications in the field of drug discovery and development.

The two parties’ choices have ultimately converged, both targeting the AI for Science sector.

Isomorphic Labs: The Biggest Bet in AI-Driven Drug Discovery

Isomorphic LabsFounded in2021By Google at the end of the year,DeepMindTransformed from the team.

The development of traditional drugs typically takes an average of ten years and costs billions of dollars, with varying success rates and methodologies for different targets.

andIsomorphic Labs is using AI to create aA universal drug design engine that can be applied not only to single targets or single drug types, but also repeatedly across any different disease areas.

May 2024,Following the tradition ofAlphaFold Approach: Isomorphic Labs and DeepMind jointly released AlphaFold 3, which can predict the structures and interactions of all biomolecules, becoming a revolutionary tool for drug discovery.

In April 2025, the company announced that it had obtained$600 millionof the massive financing, byLed by Thrive Capital, with participation from GV and Alphabet. The funding will support their AI-driven drug design engine and help advance their internal projects into clinical stages.

This YearFebruary, Isomorphic Labs launchedOutIsoDDE, is currently the world's most powerful drug designModel A.

It predicts the binding affinity of small molecules with greater accuracy than physics-based methods, and in proteins—Accuracy Ratio in Ligand Structure PredictionAlphaFold3More than doubled.Nowadays,Isomorphic Labs and Multinational Pharmaceutical CompaniesNovartisandEli LillySigned a major research collaboration, with a total amount exceeding$3 billion.

President of the CompanyMax Jaderberg stated, “Our drug is poised to enter the clinical phase, where we will begin to observe the efficacy of these molecules. This will be a very exciting moment.”

No specific details have been confirmed regarding the timeline, therapeutic approach, or target; however, it has already surpassed the previously estimated timeframe for initiating clinical trials.Late 2025.

In the future, Demis Hassabis will devote greater energy toIsomorphic, looking forward to the company's future performance.

Zhì Yào Jú will continue to monitor.

Recommended Reading