On Tuesday (November 25), NVIDIA officially stated that its technology is still one generation ahead of the industry and is the only platform that can run all artificial intelligence (AI) models and be applied to all computing scenarios. Analysts believe that Nvidia’s move is in response to Wall Street’s concerns that the company’s dominance in AI infrastructure may be threatened by Google chips.

Nvidia posted on social platform

"NVIDIA is a generation ahead of the industry - the only platform that can run all AI models and be deployed in all computing scenarios."

Nvidia added: "Compared to ASIC (application-specific integrated circuit) chips designed for specific AI frameworks or functions, Nvidia provides higher performance, greater versatility, and better substitutability."


This statement comes amid reports that Meta, one of Nvidia's important customers, may reach an agreement with Google to use Google's tensor processing units (TPUs) in data centers. Affected by this, Nvidia's stock price fell by more than 7% during the day.

Last week, Google released its latest large language model, Gemini 3, which many industry authorities believe has "surpassed" OpenAI's GPT model. The model was trained using TPUs, not NVIDIA GPUs.

Analysts pointed out that Nvidia's share of the AI ​​chip market exceeds 90%, and although Blackwell is expensive, it has powerful performance.

However, in recent weeks, Google's self-developed chips have received increasing attention as a viable alternative to Blackwell chips. Executives within Google Cloud revealed that expanding TPU market adoption is expected to help the company capture 10% of Nvidia's annual revenue share.

Unlike Nvidia, Google does not sell its TPU chips to other companies, but it uses them for internal tasks and allows businesses to rent them through Google Cloud.

A Google spokesperson mentioned in a previous statement: "Demand for both our self-developed TPUs and NVIDIA GPUs is accelerating. We will continue to support both technologies as always."

Nvidia CEO Jen-Hsun Huang also talked about competition from TPU during an earnings call earlier this month. He pointed out that Google itself is a customer of Nvidia's GPU chips and that the Gemini model can also run on Nvidia's technology.

Huang Renxun also mentioned that he has been in contact with Demis Hassabis, CEO of Google DeepMind.

Huang Jensen said Hassabis texted him and said that the technology industry’s theory that “using more chips and data can build more powerful AI models”—what AI developers often call “scaling laws”—still holds true.

NVIDIA believes that the law of scale will drive stronger demand for its chips and systems.