Google developing new AI chip 'Frozen v2' for Gemini efficiency
Original: Google is working on a new AI chip designed to make Gemini more efficient
Why This Matters
Custom AI chip development signals a critical industry shift away from Nvidia dependency toward proprietary, cost-efficient hardware.
Alphabet is developing a new server chip internally called 'Frozen v2,' targeting a 2028 release. According to The Information, the chip could be 6–10x more efficient than Google's current AI chips, measured by tokens generated per unit of power.
Alphabet is developing a new server chip codenamed 'Frozen v2,' designed to improve the efficiency of its Gemini AI models. The Information first reported the development, citing anonymous sources, and said the chip is slated for release in 2028. The chip is projected to be between six and ten times more efficient than Google's existing AI chips, with efficiency measured by the number of tokens generated per unit of power consumed.
Google did not directly confirm or deny the report to TechCrunch, stating: 'Our teams are constantly researching and experimenting with new innovations to deliver maximum performance and efficiency for our users and customers. By co-designing our hardware and software from the ground up, we ensure our systems are integrated and highly optimized for real-world workloads.'
The news comes amid a broader industry trend of AI companies developing proprietary chips to reduce dependency on Nvidia and address global shortages in AI computing capacity. In June, OpenAI announced its first custom inference chip called Jalapeño, and Anthropic was reportedly in discussions with Samsung over a chipmaking partnership. Google previously announced plans to spend between $180 billion and $190 billion on AI infrastructure. Following The Information's report, Alphabet's stock rose approximately 3% ahead of its earnings report.