Alphabet, Google’s parent company, is designing a new server chip, internally dubbed “Frozen v2,” that is aimed at making its in-house Gemini AI models run more efficiently when serving queries. According to a report from The Information, the chip is slated for release sometime in 2028 and could be between six and 10 times more efficient than Google’s existing AI chips, measured by the number of tokens generated per unit of power.
Google declined to directly confirm the report but did not deny it. “Our teams are constantly researching and experimenting with new innovations to deliver maximum performance and efficiency for our users and customers,” a Google spokesperson told TechCrunch. “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.”
What is Frozen v2 and what would it do?
Frozen v2 is being designed to serve Gemini, rather than to train it.
The shift signals a move toward more specialized AI hardware. Instead of treating Gemini as software running on general-purpose AI accelerators, Google appears to be co-designing parts of the model and the chip so that specific workloads map directly onto the silicon.
Why is Google building its own AI chips?
Frozen v2 is primarily aimed at easing Google’s internal compute crunch rather than being sold to outside customers. Several factors are pushing AI companies to bring more of their silicon in-house:
- Cost pressure has made serving-cost a more prominent selling point as concerns about AI spending cool some of the market enthusiasm seen in prior years.
- Owning custom silicon can help secure more AI compute capacity.
- Reducing dependence on Nvidia. Major AI makers have worked to diversify away from Nvidia, which has historically dominated the AI chip market.
Google’s capital plans highlight what is at risk.
How does this fit into the broader AI chip race?
Google is one of several major AI companies moving toward custom silicon. Google’s existing TPUs have long given it a degree of self-sufficiency, and Frozen v2 would extend that approach further by tailoring the hardware specifically to Gemini.
Because the chip is not expected until 2028, it is a longer-horizon bet rather than an immediate replacement for the hardware currently serving Gemini. If it lands as described, however, it could give Google a structural cost advantage on inference, the most power-intensive part of running an AI product at scale.
FAQ
What is Google’s Frozen v2 chip?
Frozen v2 is the internal name for a server chip Alphabet is designing to make its Gemini AI models more efficient when serving user queries. It is expected to arrive in 2028.
How much more efficient is Frozen v2 expected to be?
According to The Information, citing anonymous sources, Frozen v2 could be between six and 10 times more efficient than Google’s current TPUs, measured by tokens generated per unit of power.
Why is Google making its own AI chips instead of relying on Nvidia?
Google is bringing more silicon in-house to address internal compute shortages, reduce the cost of serving AI responses, and lessen its dependence on Nvidia, which has historically dominated the AI chip market.
Related coverage
Do you want help with your SEO or website development? Then Contact Us Here.
NewSunSEO
Grow in Google and AI Search
AI-SEO, local search, paid ads and websites from one team, with offices on Long Island and in Myrtle Beach.
- AI-SEO
- Local AI-SEO
- AI Consulting
- PPC
- Web Design
- Web Development
- Long Island Office(631) 320-1700
- Myrtle Beach Office(843) 492-7131
Request a Free Proposal
Tell us about your business and where you want to grow.
We will reply by email to talk through your goals and what a proposal would include.