After a slow start driven by corporate caution, the Nobel laureate argues that deep resources, custom chips, and new models like Gemini 3 are pushing the tech giant past OpenAI.
The Tide Turns: Geoffrey Hinton asserts Google is beginning to overtake OpenAI, fueled by the success of its Gemini 3 and Nano Banana Pro models.The Resource Advantage: Hinton cites Google’s internal chip production and vast data centers as critical edges, predicting “Google will win” in the long run.Honoring a Legacy: Despite Hinton leaving Google to warn about AI dangers, the company is donating $10 million to establish a chair in his name at the University of Toronto.
For three years, the narrative in Silicon Valley has been consistent: OpenAI moved fast and broke things, while Google, the slumbering giant, scrambled to catch up. According to Geoffrey Hinton, the “Godfather of AI” and a 2024 Nobel Prize winner in physics, that narrative is effectively over.
In a recent interview with Business Insider, Hinton stated that the momentum has shifted decisively. “My guess is Google will win,” Hinton declared, suggesting that the era of OpenAI’s dominance may be drawing to a close as Google leverages its massive infrastructure to reclaim the throne.
The Resurgence: Gemini 3 and The Chip Advantage
Hinton, a professor emeritus at the University of Toronto who spent a decade at Google Brain, expressed surprise not that Google is winning, but that it took them this long to do so. “I think it’s actually more surprising that it’s taken this long for Google to overtake OpenAI,” he noted.
The catalyst for this shift appears to be the release of Gemini 3, a model update that industry insiders claim has leapfrogged OpenAI’s current capabilities. Coupled with the hit release of the Nano Banana Pro image model, Google has effectively silenced critics who claimed it had lost its innovative edge.
However, the advantage isn’t just about software; it is about infrastructure. Hinton pointed out that Google’s move to manufacture its own AI chips is a “big advantage.” Reports suggest Google’s hardware prowess is so significant that the company may broker a billion-dollar deal to supply Meta with these chips.
“Google has a lot of very good researchers and obviously a lot of data and a lot of data centers,” Hinton explained. “I think that right now they’re beginning to overtake it.”
The Caution Trap: Why Google Held Back
To understand Google’s position, one must look at the broader history of AI development. As Hinton reminded observers, “Google was in the lead for a long time, right? Google invented transformers. Google had big chatbots before other people.”
If Google had the technology first, why did they lose the initial public relations war to ChatGPT? The answer, according to Hinton, lies in reputational risk. He cited the disastrous 2016 launch of Microsoft’s “Tay” chatbot, which was taken offline within 24 hours after it began posting racist tweets.
“Google, obviously, had a very good reputation and was worried about damaging it like that,” Hinton said. This sentiment was echoed by Google CEO Sundar Pichai, who admitted earlier this year that the company held back because their products “hadn’t quite gotten it to a level where you could put it out and people would’ve been okay with Google putting out that product.”
Growing Pains and Recent Stumbles
Google’s cautious approach wasn’t entirely unfounded, as evidenced by its own rocky rollouts once it decided to speed up. Just last year, the company had to pause its AI image generator following backlash over “woke” results that displayed historically inaccurate images of people of color. Furthermore, early iterations of its AI search overviews provided nonsensical and potentially dangerous advice, such as suggesting users put glue on pizza to keep the cheese in place.
Despite these stumbles, the sheer scale of Google’s “Code Red”—declared three years ago in response to ChatGPT—seems to have paid off. Recent reports indicate that the alarm bells are now ringing inside OpenAI rather than Google.
A Complicated Legacy Honored
The relationship between Hinton and Google remains one of mutual respect, despite their ideological divergence on the future of AI safety. Hinton famously left Google in 2023 to speak freely about the existential risks AI poses to humanity, from job displacement to the potential for machines to outsmart their creators.
Yet, recognizing his foundational contributions, Google announced a $10 million CAD donation to the University of Toronto to establish the “Hinton Chair in Artificial Intelligence.” The university has pledged to match this gift.
“Geoff’s work on neural networks — spanning his time in academia and his decade here at Google — laid the foundation for modern AI,” Google said in a statement. It is a fitting tribute to the man who laid the groundwork for the very technology that Google is now using to win the race he helped start.