Google Reportedly Developing New AI Chip to Boost Gemini Performance and Improve Energy Efficiency

Google is reportedly working on a next-generation artificial intelligence chip designed to enhance the performance of its Gemini AI models while significantly improving energy efficiency. The new processor, currently under development, is expected to strengthen Google's AI infrastructure and support future advancements in generative AI.

According to a report by The Information, the chip—internally codenamed Frozen V2—is targeted for manufacturing in 2028 and is being developed as a dedicated AI accelerator for data center workloads.

Google's upcoming AI chip is expected to enhance the performance of its AI models while strengthening its position in the global AI race

Designed for Faster and More Efficient AI

The upcoming processor is expected to work alongside Google's existing Tensor Processing Units (TPUs), helping AI systems process complex workloads more efficiently.

Reports suggest the new chip could generate six to ten times more AI tokens using the same amount of electricity compared with Google's current AI hardware. If achieved, this would allow Gemini models to deliver faster responses while reducing energy consumption and operational costs.

The efficiency gains could also improve scalability as demand for AI-powered services continues to grow.

Google Focuses on Hardware-Software Integration

Responding to reports, a Google spokesperson said the company continues to explore new hardware innovations to maximize AI performance and efficiency.

The spokesperson noted that Google's strategy centers on developing hardware and software together, allowing its systems to be optimized for real-world AI applications and enterprise workloads.

While not every research project ultimately reaches production, the company said ongoing experimentation remains a key part of its long-term AI development roadmap.

Competition Intensifies in AI Chip Development

Custom AI chips are becoming increasingly important as technology companies seek greater control over computing performance and infrastructure costs.

Google's reported plans come as major AI firms continue investing in proprietary hardware. Companies including OpenAI, Microsoft, Meta, and Amazon are expanding their AI infrastructure to support increasingly powerful models.

Industry reports also indicate that OpenAI recently introduced its own custom inference processor, while Anthropic is exploring partnerships to strengthen its chip development capabilities.

Gemini Development Continues

The report also arrives amid speculation surrounding the launch timeline of Google's next flagship AI model, Gemini 3.5 Pro.

According to reports, the company is continuing to refine and train the model before its public release, with efforts focused on improving overall performance and reliability.

If the new AI chip reaches production as planned, it could play a significant role in supporting future generations of Gemini and enhancing Google's competitiveness in the rapidly evolving AI industry.