The Washington Post/Getty
There are plenty of stories out there about just how ruthless the hiring wars for AI talent have gotten.
Perplexity's CEO recounted earlier this year how he was rebuffed by a Meta engineer who told him to "come back to me when you have 10,000 H100 GPUs," the highly sought-after Nvidia chips needed for cutting-edge AI development.
But an AI pioneer's rehiring by Google might just take the cake.
Google paid an eye-popping $2.7 billion for a deal that brought 48-year-old Character.ai founder Noam Shazeer back into the fold, according to a recent report by The Wall Street Journal.
Now, that didn't all go to Shazeer — Google didn't formally acquire the company but paid to license the startup's technology — but the AI engineer made hundreds of millions, the Journal reported, citing a source familiar with the matter.
But the main play was to convince Shazeer to work at Google again, according to the report, and he's now helping lead the next version of its Gemini AI. Google did not respond to a request for comment from Business Insider.
While it's certainly one of the most notable examples to date of the sky-high price that tech companies are willing to pay for AI talent, Shazeer isn't the average AI engineer. He co-authored a famous research paper with seven other Googlers in 2017 called "Attention is All You Need," which helped launch the generative AI boom.
He then left Google in 2021 after over two decades at the company when the tech giant refused to release his chatbot over concerns it might say the wrong thing.
But a lot can happen in three years. And as AI competition heats up, tech companies are engaged in fierce competition to develop the best Similar to the deal Google made, Microsoft previously paid Inflection $650 million for a licensing deal and to poach its founders Mustafa Suleyman and Karen Simonyan.
Andree Mendoza, a technical recruiter at talent firm Carex Consulting Group, told Business Insider that there will likely be more deals like this moving forward. Mendoza said these strategic licensing agreements are a "clever way to reel in top talent without jumping through all the hoops of a full acquisition."
And beyond cutting huge checks, Big Tech firms are pulling in their CEOs and founders to help close the deal — Google cofounder Sergey Brin reportedly helped convince Shazeer to boomerang back. Zuckerberg has personally written emails to Google DeepMind AI researchers in an effort to recruit them, The Information reported. The report also said Meta has even offered jobs to candidates without interviewing them.
OpenAI CEO Altman has used similar tactics, calling applicants to persuade them to join the company. As of June, the ChatGPT maker had hired at least 44 ex-Googlers.
While these grand efforts and massive payouts may seem like a bit much, Conor Grennan, Chief AI architect at NYU's Stern School of Business, told Business Insider that it's "money well spent." The AI industry is more like professional sports right now than a typical business model, Grennan said.
"Noam Shazeer isn't just an employee; he's a star quarterback. There are simply very few individuals like him in the world," Grennan said, adding that few have the intellect and track record to successfully create a company like Character.ai.
The platform had about 20 million active users a month, according to The Wall Street Journal report. It allows users to talk to fictional AI characters and AI celebrities.
Mendoza, who also compared the AI industry to professional sports, said that the market for high-end and specialized AI talent is small and only a handful of people have the mix of "deep technical expertise" within a large enterprise.
"A lot of the experts are already with the big players or have launched their own startups," Mendoza said.
Grennan said even if Google did overpay, that's the nature of the game — and companies like Google, Meta, Apple, and OpenAI can afford it.
"And if just one of these talented individuals 'hits,' that's a product worth billions," Grennan said.
Grennan predicted this pattern will continue as the AI talent window stays hot for the next three to four years. But eventually, he expects it to shrink dramatically as AI becomes more capable and fewer people are needed to build it out.
