Google's DeepMind Crumbles Under Pressure of Delays and Employee Exhaustion
Internal tensions at Google's DeepMind have been building for months, according to current and former employees. Stalled AI models, missed project deadlines, and widespread staff burnout are cited as key factors behind the unit's decline. The situation highlights the human cost of high-stakes research environments.
The leadership restructuring moves operational control to Mountain View, with Koray Kavukcuoglu reporting directly to Sundar Pichai rather than assuming a standalone DeepMind CEO role. London-based staff perceive this as a gradual erosion of the lab's independent identity, with some fearing Demis Hassabis's chairman position is a precursor to his complete exit.
Underlying the reshuffle, the lab's flagship Gemini 3.5 Pro model has slipped past three release deadlines, which engineers attribute to a misallocation of focus away from coding capabilities. Combined with a public dispute over defense contracts and the departure of prominent researchers like Jeff Dean, these pressures have fueled internal burnout and a sense of falling behind rivals OpenAI and Anthropic.
This internal upheaval could signal a broader reckoning for high-stakes AI research. If burnout and talent attrition persist at DeepMind, the pace of AI advancement may slow, affecting businesses and consumers dependent on cutting-edge models. The structural shift toward corporate control might also deter independent-minded researchers, potentially concentrating AI power within a few dominant firms. For tech employees, this episode may highlight the unsustainable human costs of relentless competition, prompting industry-wide reassessment of workload and retention strategies.