Google is reportedly working on a next-generation artificial intelligence chip that could significantly improve the performance and efficiency of its Gemini AI models.

According to recent industry reports, the new processor, internally known as “Frozen v2,” is expected to launch around 2028. The chip is designed to process AI tasks using much less power while generating responses much faster than Google’s current AI hardware.

Although Google has not officially confirmed the project, the company says it continues to invest heavily in developing new hardware and software together to improve AI performance.

What Is Frozen v2?

Frozen v2 is believed to be Google’s upcoming custom AI accelerator built specifically for running Gemini models more efficiently.

Reports suggest the chip could deliver six to ten times better efficiency compared to Google’s existing AI processors. Instead of simply increasing raw computing power, the new design focuses on generating more AI output while consuming less electricity.

This could help Google reduce operating costs while making Gemini faster for millions of users.

Why Google Is Building Its Own AI Chips

Artificial intelligence requires enormous computing power.

Every AI request—from generating text to creating images or analyzing documents—depends on specialized processors inside massive data centers.

Until recently, most AI companies relied heavily on chips from NVIDIA, which currently dominates the AI hardware market.

However, major technology companies are increasingly designing their own chips to:

  • Reduce dependence on third-party suppliers
  • Lower AI operating costs
  • Improve energy efficiency
  • Optimize hardware specifically for their own AI models

Google has already developed several custom Tensor Processing Units (TPUs), and Frozen v2 appears to be the next step in that strategy.

AI Hardware Competition Is Heating Up

Google is not the only company investing in custom AI processors.

Several major AI companies are building their own hardware to improve performance and reduce infrastructure costs.

Recent developments include:

  • OpenAI introducing its first custom inference chip.
  • Anthropic reportedly exploring semiconductor partnerships.
  • Microsoft, Amazon, and Meta continuing to expand their own AI infrastructure.

As AI adoption grows worldwide, efficient hardware is becoming just as important as advanced AI models.

Investors Are Watching Google’s AI Investments

Google has committed billions of dollars toward expanding its AI infrastructure.

The company previously announced plans to spend heavily on AI development, including cloud infrastructure, data centers, and custom hardware.

Reports about the new Frozen v2 chip appeared to boost investor confidence, with Alphabet shares rising following the news.

Investors see custom AI chips as an important way to reduce long-term costs while improving Gemini’s competitiveness against rival AI platforms.

Google Hasn’t Confirmed the Details

When asked about the reports, Google did not directly confirm the existence of Frozen v2.

Instead, the company stated that its engineering teams are constantly researching new technologies and experimenting with hardware innovations designed to improve performance and efficiency.

Google also emphasized its long-term strategy of designing hardware and software together to deliver better AI experiences for users and businesses.

Stay updated with the VitalStack.

Read More on VitalStack

Enjoyed this article?

Subscribe for weekly deep-dives on AI and health — straight to your inbox.