Summary of Key Events
Google has experienced two significant personnel changes in one day: Demis Hassabis, the founder of DeepMind, resigned from his role as CEO to focus on long-term AI research, while legendary engineer Jeff Dean left the company to start his own business (focusing on AI automation research). These moves indicate a shift in Google's priorities from emphasizing basic AI research to focusing more on product development and commercialization in order to compete with rivals such as OpenAI and Anthropic. Upon the announcement of these changes, Google's parent company, Alphabet, saw its stock price plummet by 4%, resulting in a loss of over $180 billion in market value. However, behind these changes lies Google's strategic adjustment to balance long-term research with short-term commercial goals, while also revealing challenges such as talent loss, lagging products, and financial pressure.
1. The Personnel Changes Are Not Just “Leavings,” but a “Shift in Strategic Focus”
The departures of these two key figures are not mere resignations; they represent a reorganization within Google's AI strategy for the new competitive landscape:
- Demis Hassabis: Moving from Manager to Scientist
Previously the CEO of DeepMind, he now serves as the chairman and chief scientist at Alphabet, freeing up his time to focus on long-term research projects such as general artificial intelligence (AGI) and scientific breakthroughs. While DeepMind could spend years on developments like AlphaGo and AlphaFold, the industry demands faster product releases. Therefore, more execution-oriented figures like Dario Kavukcuoglu (the former CTO) have taken over to commercialize Gemini models, aiming to sell them to companies or integrate them into Google’s search and cloud services.
- Jeff Dean: Starting a Business with Google’s Support
Known as the “god of engineering” for his exceptional coding skills, Dean has left Google to co-found Discovery Loop with several other researchers. Their goal is to automate the entire scientific research process using AI (e.g., automatically designing experiments and analyzing data). Google continues to support them, reflecting a new approach in the AI era: rather than completely losing these talents, the company is collaborating to explore future directions while still maintaining the potential for long-term research.
2. Why These Two Are Considered “The Soul of Google”? Their Departures Mark the End of an Era
These individuals are not just ordinary employees; they embody Google’s technical expertise and cultural values:
- Jeff Dean: The Pillar of Google’s Engineering Tradition
Joining Google in 1999, he contributed to the invention of MapReduce (a technology for processing large datasets) and Bigtable (the core system for data storage). He also played a key role in establishing Google Brain and developing TensorFlow, a tool widely used by AI developers worldwide. His work reflects Google’s commitment to excellence through meticulous engineering.
- Demis Hassabis: Leading Google from a “Mobile First” Approach to an “AI First” Strategy
In 2014, he led DeepMind, which was acquired by Google with only 20 employees and no products. He later made AI popular globally with AlphaGo and solved the scientific challenge of protein structure prediction with AlphaFold. His efforts shifted Google’s focus from revenue generation through search and Android to investing in the future of AI.
3. Rivals Are Moving Too Fast, Forcing Google to Accelerate
Google was once a leader in the AI field, but now it is being chased by competitors like OpenAI and Anthropic:
- Significant Gap in Product Iteration Speed
While rivals are updating their models on a weekly basis (e.g., Anthropic’s Claude Opus 5 and OpenAI’s GPT-5.6), Google’s Gemini 3.5 Pro has been delayed, even dropping out of the top ten in rankings for programming intelligence agents.
- Lagging in Commercial Revenue
In 2026, Anthropic’s annual revenue was $62 billion, and OpenAI’s was close to $60 billion—combined, they exceed $100 billion. Google’s AI revenue mainly comes from selling computing power (through its cloud services), with much less income from pure model APIs and applications compared to its competitors.
- Inefficient Internal Processes
Google’s lengthy approval processes and strict security reviews have hindered innovation, as DeepMind was not even allowed to publish papers freely. With the rapid pace of the industry, Google must address these inefficiencies.
4. Google Faces Additional Challenges
Even after these adjustments, Google still faces three major problems:
- Talent Loss
Nobel laureate John Jumper has joined Anthropic, and the author of Transformer (a fundamental AI technology) has gone to OpenAI. There are concerns about whether Google can continue to attract top AI researchers.
- Pressure to Catch Up with Competitors
Although Hassabis mentions the advancement of Gemini 4, the delay of Gemini 3.5 has shaken market confidence. The new CEO, Dario Kavukcuoglu, needs to quickly deliver products that can compete with GPT-5 and Claude 5.
- Financial Strains
To build its AI infrastructure (chips, data centers), Google spent $44.9 billion in the second quarter, resulting in a negative free cash flow of -$5.9 billion. With annual expenses expected to range between $195 billion and $205 billion, the company is facing significant financial pressure.
In Conclusion
Google’s personnel changes are a result of being forced to adapt: it is shifting from a focus on scientists to a balance of engineers and product managers in order to keep up with the competition in AI. However, success will depend on its ability to retain top talent, release products quickly, and strike a healthy balance between long-term research and short-term profitability. In the AI era, speed is everything.