Google pays AI technicians in the industry’s talent war

The report cites four former employees of DeepMind, a subsidiary of Google AI, who said they were placed on what is called “extended garden leave.” Under this arrangement, engineers stopped working for Google, but continued to receive salaries – condition that they did not work with any competing AI companies during the non-competition clause.
Although non-competitive protocols are common in the technology field to protect know-how, the degree to which they are restricted on Google DeepMind seems unusual. Typically, DeepMind’s non-competitive period lasts about six months, but at least two engineers interviewed claimed that their vacation had been extended by a full 12 months, especially for engineers related to the Gemini AI project – Google’s flagship AI initiative.
This approach, while not unprecedented, marks Google’s bold move to secure its competitive advantage in a rapidly growing landscape. With competitors such as Meta and Openai accelerated innovation, the company is trapped in a fierce competition for AI supremacy. Google’s own release of Gemini 2.5 Pro (touted as the most advanced AI model to date) is part of this escalating competition.
However, the strategy sparked debate. Although Google believes its contracts are “compliant with market standards”, critics believe that such an expanded leaf is too restrictive and may kill innovation. Some former DeepMind employees described the policy as excessive and effectively distanced valuable talents as a result of a surge in demand for AI expertise.
As artificial intelligence continues to shape the future of technology, the struggle for talent is intensifying. In this high-risk game, companies are not only competing in products, but also competing with people. In Google’s case, this means staying top talent on the payroll even if they don’t work.