• 2 min read
Google’s Frozen v2 chip targets 10x Gemini efficiency
Google is reportedly developing Frozen v2, a 2028 server chip designed to make Gemini six to 10 times more power-efficient.
Image: TechCrunch
Alphabet is reportedly developing a new server chip that could make its Gemini models six to 10 times more power-efficient than the company’s current AI hardware.
The chip, internally known as “Frozen v2,” is scheduled for release sometime in 2028, according to The Information, which cited anonymous sources. Its efficiency is measured by the number of tokens generated per unit of power.
Google did not directly confirm or deny the report in a response to TechCrunch. The company said its hardware teams are continuing to explore new designs, while warning that not every project reaches production.
“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.”
Google’s push for custom AI chips
AI companies are increasingly designing their own processors to improve the efficiency of their models and reduce their reliance on scarce AI computing capacity. Efficiency has become a more prominent selling point as concerns over AI spending have cooled some of the market excitement around the sector.
The effort also reflects an industry-wide attempt to reduce dependence on Nvidia, whose dominant position in AI chips has left major model developers reliant on its hardware. OpenAI announced its first custom chip, an inference processor called Jalapeño, in June. Earlier this month, reports said Anthropic was discussing a chipmaking partnership with Samsung.
Recommended reading
Roscosmos tests AI for ISS crew reports
Investors have also questioned Alphabet’s planned AI spending. Earlier this year, Google said it expected to spend between $180 billion and $190 billion as it builds out its AI strategy.
The Frozen v2 report appeared to reassure investors ahead of Alphabet’s earnings report later this week. Google’s stock rose by about 3% on Monday morning after The Information published its report.
AI Editor
Ava covers the rapidly evolving world of artificial intelligence, from foundational models and research labs to the real-world economics of intelligence. With a background in computational linguistics, she cuts through the hype to find out what actually works. She firmly believes that benchmarks are just marketing until reproduced in the wild.
via TechCrunch
