Google’s parent company Alphabet is reportedly designing a new server chip aimed at making its in-house Gemini models run more efficiently.
The chip, internally called “Frozen v2,” is expected to release
sometime in 2028, according to a report by The Information, citing anonymous sources. The report claims the chip could be between six and ten times more efficient than Google’s current AI chips, measured by the number of tokens generated per unit of power.
Google didn’t directly confirm or deny the report when TechCrunch reached out. In its response, the company said: “Our teams are constantly researching and experimenting with new innovations to deliver maximum performance and efficiency for our users and customers. While not every project moves into production, this rigorous exploration is central to our full stack approach. By co-designing our hardware and software from the ground up, we ensure our systems are integrated and highly optimized for real-world workloads.”
AI companies have increasingly moved towards building their own chips, both to make their models run more efficiently and to work around ongoing shortages in AI computing capacity.
Efficiency has become a genuine selling point for tech companies lately, especially as concerns over AI spending have cooled some of the market enthusiasm that previously surrounded the industry.
At the same time, several major AI firms are trying to reduce their reliance on Nvidia, which has long dominated the AI chip market and left many companies dependent on its hardware.
This trend isn’t limited to Google. In June, OpenAI announced its first custom chip, an inference processor called Jalapeño. Earlier this month, reports emerged that Anthropic was in talks with Samsung on a new chipmaking partnership.
Investors have previously raised concerns over Alphabet’s massive planned spending tied to its AI strategy. Earlier this year, Google said it plans to spend between $180 billion and $190 billion on this front.
With that much money on the line, the company has real incentive to show these investments are paying off, and a more efficient in-house chip would be a concrete way to make that case.
