Google overtakes OpenAI, will win race
The “Godfather of AI” Geoffrey Hinton sees a decisive turning point in Google’s favor in the AI race. According to the Nobel Prize winner, thanks to Gemini 3 and its own TPUs, the company has now technically overtaken the former pioneer OpenAI.
Technological leadership change in the AI sector
Geoffrey Hinton, often referred to as the “Godfather of AI” and recipient of the Nobel Prize in Physics in 2024, is observing a significant shift in the balance of power in the field of artificial intelligence. While OpenAI was long considered the undisputed leader with the chatbot ChatGPT, Hinton believes Google has now taken the lead. The decisive factors for this are the successful launch of Gemini 3, which is seen in the industry as a direct and powerful answer to OpenAI’s GPT-5, as well as the presentation of new, advanced image models. But Hinton is surprised that it took the search engine giant so long to achieve this dominant position, since the financial and technological resources were always there.
The tide seems to have turned three years after the “Code Red” that Google management declared shortly after the release of ChatGPT, because the alarm bells are currently ringing among the competition. Google has massive amounts of data from search, YouTube and other services, as well as an infrastructure that has grown over decades. The company is now finally using these advantages effectively to leave the competition behind and integrate products more deeply into its own ecosystem. Google is proclaiming a “new era of intelligence” with Gemini 3
Own chips as a strategic advantage
A key factor in this change is the group’s hardware strategy. As Geoffrey Hinton said in an interview with Business Insider explained, Google has an immense strategic advantage over competitors who rely on third-party hardware thanks to its long-term development of its own AI chips – the so-called Tensor Processing Units (TPUs). “Google has a lot of great researchers and obviously a lot of data and a lot of data centers,” the former Google Brain researcher told the magazine, adding, “I suspect Google will win.” This strength is underpinned by current reports that a deal worth a billion dollars is in the pipeline to supply the Meta Group with these same AI chips.
The fact that Hinton is even talking about a catching-up process is due to Google’s previous reluctance. The company itself invented the Transformer architecture on which modern language models are based and had powerful chatbots like LaMDA early on, but deliberately kept these under wraps. According to Hinton, the reason for this lies in the past of the competition: the debacle surrounding Microsoft’s chatbot “Tay” in 2016, which published racist tweets within 24 hours, served as a deterrent. Google didn’t want to jeopardize its good reputation with immature products. This caution was well founded, as later mishaps showed.
An early version of the AI-based “AI Overviews” in Search advised users to put non-toxic glue on pizza to prevent the cheese from slipping – advice that was based on a satirical Reddit comment. Gemini’s image generator also had to be temporarily paused after it produced historically incorrect representations, such as various Vikings or Wehrmacht soldiers, which critics described as forced “woke.” However, the company has largely overcome these teething problems and is now scaling its models at a speed that impresses even critics.
Digital marketing enthusiast and industry professional in Digital technologies, Technology News, Mobile phones, software, gadgets with vast experience in the tech industry, I have a keen interest in technology, News breaking.
