With the AI talent wars heating up between companies like OpenAI, Meta and Anthropic, one way Google has been competing is by aggressively rehiring former employees.

Some 20% of software engineers working on artificial intelligence that Google hired in 2025 were so-called boomerang employees, an increase from prior years, CNBC has learned. A Google spokesperson confirmed the statistic remains accurate as of December, and said the company saw a jump in the number of AI researchers coming from major competitors compared to 2024.

“We’re energized by our momentum, compute, and talent — engineers want to work here to keep building groundbreaking products,” the spokesperson said in a statement.

John Casey, Google’s head of compensation, recently told employees in a meeting about the rehiring. Casey said AI-focused software engineers are drawn to Google’s deep pockets and hefty computational infrastructure that’s needed to perform advanced AI work, according to audio reviewed by CNBC.

Google has a large pool of ex-employees to mine, particularly after its largest ever round of layoffs in early 2023, when parent company Alphabet cut 12,000 jobs, reducing headcount by 6%. That followed a market downturn driven by soaring inflation and rising interest rates. Google has since continued with rolling layoffs and buyouts.

Across the industry, employee boomerangs are up, according to data published earlier this year by ADP Research, with the sector it classifies as information showing the starkest numbers.

Google has been racing to catch up in generative AI after a slow start that followed OpenAI’s release of ChatGPT in late 2022. After fumbling a number of product rollouts, the company has bounced back this year, thanks to hefty investments in AI infrastructure and the success of its Gemini app. Google announced its latest model, Gemini 3, last month.

Alphabet’s stock price is up more than 60% this year, outperforming all of its megacap peers.

As a historical hotbed of engineering and innovation, Google has long been a place where competitors have turned to try and poach talent. That’s still the case.

Earlier this year, Microsoft hired around two dozen employees from Google’s DeepMind AI research lab, CNBC reported in July. OpenAI, meanwhile, has opened its wallets wide, along with Meta. OpenAI CEO Sam Altman told employees in June that Meta had been offering $100 million signing bonuses, and that he was aggressively trying to retain staffers.

Late last year, Google brought back a major figure in AI: Noam Shazeer.

Shazeer and Daniel De Freitas left Google in 2021 to start AI platform Character.AI, reportedly departing after Google rebuffed their attempts to try and get the company to push its internal chatbot forward.

Along with other members of the Character.AI research team, Shazeer and De Freitas rejoined DeepMind in August 2024 under a licensing deal for the startup’s technology.

Over the last year, Google has taken more risks, shipping products more quickly, even if they aren’t viewed as completely ready. Google has also made a companywide effort to remove bureaucracy, enacting widespread employee buyouts and eliminating more than one-third of its managers overseeing small teams, CNBC reported in August.

Google co-founder Sergey Brin, who came out of retirement in 2023, has at times personally reached out to prospective candidates to recruit them, according to people familiar with the matter who asked not to be named because they weren’t authorized to speak to the media. Meta CEO Mark Zuckerberg has also reportedly reached out to researchers on behalf of his company.

WATCH: OpenAI’s Sam Altman says Google is still a huge threat

Palo Alto Networks will migrate key internal workloads to Google Cloud as part of a new multibillion-dollar agreement, the companies announced on Friday.

The companies said the deal is an expansion of their existing strategic partnership and will deepen their engineering collaboration.

Palo Alto Networks is now using Google’s Gemini artificial intelligence models to power its copilots, and it is also using Google Cloud’s Vertex AI platform, according to a release.

“Every board is asking how to harness AI’s power without exposing the business to new threats,” BJ Jenkins, president of Palo Alto Networks, said in a statement. “This partnership answers that question.”

Palo Alto Networks, which offers a range of cybersecurity products, already has more than 75 joint integrations with Google Cloud and has completed $2 billion in sales through the Google Cloud Marketplace.

As part of the new phase of the partnership, Palo Alto Networks customers will be able to protect live AI workloads and data on Google Cloud, maintain security policies, accelerate Google Cloud adoption and simplify and unify their security solutions, the companies said.

Shares of Palo Alto Networks were up 1% on Friday. Google shares were mostly flat.

“This latest expansion of our partnership will ensure that our joint customers have access to the right solutions to secure their most critical AI infrastructure and develop new AI agents with security built in from the start,” Google Cloud President Matt Renner said in a statement.

WATCH: Google unveils ‘Gemini 3 Flash’ AI model focused on speed and cost