In a surprising twist in the cutthroat world of artificial intelligence (AI), Google’s AI division, DeepMind, is allegedly paying select staff in the U.K. to do nothing for up to a year rather than letting them join competitors like OpenAI or Anthropic.
The move, first reported by Business Insider, highlights the lengths to which tech giants will retain top talent in the rapidly evolving AI sector.
According to sources familiar with the situation, DeepMind has implemented “aggressive” noncompete agreements for certain employees at its London headquarters.
These contracts reportedly prevent staff from working for rival firms for 12 months after leaving Google, while still compensating them during that period.
The strategy appears designed to keep valuable AI expertise out of the hands of competitors amid an intensifying race to dominate the field.
Google, which acquired DeepMind in 2014 for roughly $500 million, told Business Insider that it uses noncompetes “selectively.”
The company’s approach contrasts with recent U.S. regulatory moves, where the Federal Trade Commission banned most noncompete clauses in 2024.
However, that ruling does not apply to DeepMind’s U.K. operations, where such agreements remain legally permissible under U.K. employment law, though controversial.
A Talent War in AI
The AI industry is witnessing unprecedented competition as companies vie for a shrinking pool of skilled researchers and engineers.
With breakthroughs in generative AI, machine learning, and neural networks driving innovations—from chatbots to autonomous systems—retaining talent has become a strategic priority.
DeepMind has long been a leader in AI research, producing milestones like AlphaGo, which defeated world champion Go player Lee Sedol in 2016.
Yet, the rise of rivals like OpenAI—backed by Microsoft and responsible for ChatGPT—and Anthropic, founded by ex-OpenAI researchers with funding from Google and Amazon, has intensified the pressure.
Posts on X reflect growing chatter about the talent scramble, with some users speculating that Google’s strategy could backfire by alienating employees eager to work on cutting-edge projects elsewhere.
“Paying people to sit on the sidelines might keep them from rivals, but it’s a risky bet,” said Dr. Emily Carter, an AI policy expert at the University of Oxford.
“These are highly motivated individuals who thrive on solving complex problems. A year of inactivity could push them to leave the field entirely—or jump ship the moment those contracts expire.”
The Cost of Inaction
For Google, the stakes are high. DeepMind’s work has bolstered the company’s AI ambitions, including enhancements to Google Search and the development of its Gemini model, which CEO Sundar Pichai has called a “critical” focus for 2025.
Losing talent to competitors could slow progress at a time when Google faces criticism for lagging in the generative AI race, sparked by OpenAI’s ChatGPT debut in 2022.
The financial burden of these noncompete payouts remains unclear, but industry analysts estimate that compensating even a small cohort of senior researchers could cost millions annually.
For context, top AI talent in the U.K. commands salaries upwards of £200,000 ($260,000 USD) per year, according to Glassdoor data, excluding bonuses and equity.
A Broader Trend?
Google’s tactic may signal a broader shift in how tech giants wield their resources to maintain dominance.
In the U.S., where noncompetes are increasingly restricted, companies like Microsoft have used cloud credits and strategic partnerships to tether AI startups like OpenAI to their ecosystems, as noted in a Microsoft blog post.
Amazon, meanwhile, has poured up to $8 billion into Anthropic, securing influence without direct control, per a Reuters report.
The practice has drawn scrutiny from employees and ethicists alike. Nando de Freitas, a former DeepMind researcher, has hinted at discontent within the ranks, referencing staff reaching out “in despair” over restrictive contracts in a public X post. “It’s a gilded cage,” one anonymous ex-employee told Business Insider, describing the arrangement as lucrative but stifling.
What’s Next?
As AI continues to reshape industries, the battle for talent shows no signs of cooling. Google’s DeepMind may view its noncompete strategy as a necessary defense, but it risks sparking a backlash from workers and regulators.
In the U.K., where employment laws are under review, pressure is mounting to curb such practices, with labor unions like Unite calling them “exploitative.”
For now, the spotlight remains on Google as it navigates this high-stakes game. Whether paying staff to sit idle proves a masterstroke or a misstep, one thing is clear: in the race to lead AI innovation, no one is playing nice.
