Nvidia affirms its possession of superior technology in the context of investors worrying that its position will be threatened by Google with its new AI chip.
“We are delighted with Google’s success, they have made great progress in AI and we will continue to provide services to them. Nvidia is a generation ahead of the industry. It is the only platform that runs every AI model and works everywhere there is computing,” Nvidia announced on X on November 26.
The message came when Nvidia shares fell 3% on November 25, after information emerged that Meta, one of the company’s key customers, could reach a multi-billion dollar deal with Google to use tensor processing units (TPUs) for its data centers.
Nvidia logo at Computex 2024 exhibition in Taiwan. Image: Khuong Nha
In the X post, Nvidia asserted that its chips are more flexible and powerful than ASIC chips – integrated circuits designed for specific companies or functions, for example Google’s TPU.
Analysts say that Nvidia accounts for more than 90% of the AI chip market share with graphics processing unit (GPU) products. However, in the past few weeks, Google’s self-developed chip has attracted a lot of attention and is considered a viable alternative to Blackwell – Nvidia’s latest, powerful and expensive chip line.
Unlike Nvidia, Google does not sell TPU but uses it internally and allows other companies to rent it through Google Cloud. They also just released Gemini 3, their highly regarded advanced AI model, trained using TPU instead of Nvidia’s GPU.
“We’re seeing increased demand for both custom TPUs and Nvidia GPUs,” a Google spokesperson said. “We are committed to supporting both, as we have been for many years.”
Nvidia CEO Jensen Huang mentioned growing competition from TPUs in an earnings call earlier this month, noting Google is a GPU customer and that Gemini can run on Nvidia’s technology. He also said he contacted Demis Hassabis, CEO of Google DeepMind, and Hassabis supports the technology industry’s “law of scaling”, which holds that using more chips and data will create more powerful AI models. According to Huang, expansion will lead to even higher demand for the company’s chips and systems.