Paid to do nothing? Google’s strategic move to bench talent for a year

At a time when competition for AI talent is intensifying, Google DeepMind is using a contractual strategy designed to slow talent loss to retain its edge: keeping some of its employees on paid leave for up to a year after they leave the company, to stop them from joining competitors.

According to a recent Business Insider report, several former DeepMind staff in the UK have confirmed that these extended leaves — enforced through noncompete clauses — are becoming increasingly common within the organisation’s UK offices. During this period — often referred to as “garden leave” — these employees are still on the payroll, even though they no longer contribute to ongoing projects.

Not all DeepMind noncompetes are created equal

The length of these noncompete periods typically varies. For individual contributors, especially those working on high-profile projects like Google’s Gemini models, six months is standard. Senior researchers, however, may face restrictions of up to 12 months.

Google says it uses noncompetes selectively and in line with market norms. But in an AI sector that’s accelerating rapidly — where six months can see the rise of a new product or even a new company — these clauses can make it challenging for researchers to maintain momentum in their careers.

Jurisdiction plays a key role in enforcement

One factor enabling this strategy is the UK’s legal stance on noncompetes. Unlike in California — where most noncompete clauses are unenforceable — UK law allows them if they’re considered reasonable and designed to protect a company’s legitimate business interests.

This legal contrast has led some former employees to consider relocating to the U.S. to escape these contractual limitations. It’s a sign of how employment laws are now part of broader strategic decisions in a high-stakes industry.

Concerns over noncompetes draw attention

While being paid to wait might sound appealing at first, several former DeepMind staff say the reality is more complicated. With AI moving as fast as it is, even a short pause can mean missing out on opportunities at rival firms or fast-growing startups that may not be able to wait six to twelve months for new hires to become available.

Microsoft’s Nando de Freitas, a former DeepMind director, recently voiced concern over the issue on social media, claiming many current employees had reached out privately for advice. His post reflected a growing frustration among researchers who feel stuck between professional ambition and contractual obligation.

A reflection of AI’s rising stakes

Google’s approach speaks to a broader shift in how companies are protecting their competitive edge in AI. The technology may be evolving quickly, but the people behind it are now just as valuable — and just as fiercely protected.

As new players and technologies emerge almost monthly, the fight for talent isn’t just about who offers the best job. It’s also about who’s able — and allowed — to take it.