When Google wanted Noam Shazeer back badly enough to spend $2.7 billion getting him, it seemed like a bet the company had won. Less than two years later, he has left this time for OpenAI, the very firm Google has been trying hardest to beat.
Shazeer announced his departure on Wednesday in a short post on the social media platform X, saying he was looking forward to joining OpenAI and working with the team there. He added that leaving had not been easy and that he was proud of everything he and his colleagues had built together at Google. It was the kind of measured statement senior figures tend to make when they know the industry will be watching every word.
OpenAI’s chief executive, Sam Altman, wasted little time responding. He wrote that Shazeer was among the people he had most hoped to work with since the company’s earliest days, and that it had taken a decade but he believed the wait would prove worthwhile. It was a pointed remark, given that OpenAI and Google have spent years jostling for ground in the same space.
To understand why this departure matters so much, it helps to know something of what Shazeer actually did. He first joined Google in the year 2000, when the company was still finding its feet. Over the years that followed, he worked on some of its core systems, including an early overhaul of the spelling correction tool that millions of people use without thinking about it. But the work that made his name came later.
In 2017, Shazeer was one of eight researchers who published a paper with the dry title “Attention Is All You Need.” It introduced what became known as the transformer, a new way of designing the kind of computer programmes that process language. That paper is now considered one of the most consequential pieces of research in the recent history of computing. Every major language tool in wide use today, from ChatGPT to Google’s own Gemini to Claude, was built on the foundations it laid.
After that, Shazeer grew frustrated with what he saw as Google’s reluctance to put its more ambitious projects in front of the public. He left in 2021 and co-founded Character.AI, a company that let ordinary users build and talk to customised chatbots. It proved popular and grew quickly. When Google decided it wanted Shazeer back, it did so through a licensing deal with Character.AI worth around $2.7 billion, described at the time as one of the most expensive arrangements of its kind ever struck in the technology industry. Shazeer returned to Google as a vice president of engineering, taking on a senior role helping to steer the Gemini programme.
The timing of his exit is notable. OpenAI is widely reported to be preparing to list its shares on the public markets, a step that could value the company at around a trillion dollars. Before any such listing takes place, companies tend to want their senior researchers signed up and settled, rather than looking elsewhere. Bringing in someone of Shazeer’s standing just before that process gets under way sends a signal about where the talent in this field currently thinks the momentum lies.
For Google, the loss arrives at a moment when the company had, by most accounts, been narrowing the gap with OpenAI’s products. Gemini had improved considerably over the past year, and Shazeer had been recognised both inside and outside the company as a significant part of that progress. Google’s response to the news was brief, telling the news agency Reuters that it was grateful for his contributions. No detail was offered about the circumstances, and nothing was said about what might happen next.
What Shazeer will actually do at OpenAI has not been spelled out publicly. Neither his title nor the scope of his role had been announced by the time the news broke. What is clear is that he arrives carrying a reputation built over more than two decades, and that his move will be read across the industry as something more than a standard change of employer.
The competition to attract and hold on to the best researchers in this field has become as intense as the competition to build the best products. Google’s experience with Shazeer, spending heavily to bring him back only to see him leave again within two years, illustrates how little money alone can guarantee in a sector where the most sought-after people have no shortage of options.
About the Author
Marcel Chidozie is a tech analyst and writer covering foreign news, fintech, and emerging technologies at TechRegard. Based in Nigeria, He's passionate about translating complex tech developments into compelling, accessible stories for diverse audiences. His work focuses on how technology shapes innovation across Africa and globally.
