Alphabet, Google’s parent company, is developing a new server chip, known internally as “Frozen v2,” to enhance the efficiency of its Gemini AI models. The release is slated for 2028, as reported by The Information, citing anonymous sources. The Frozen v2 chip is projected to be six to ten times more efficient than Google’s current AI chips, based on power usage per token generated.
Google did not confirm or deny the report when contacted by TechCrunch. The company stated, “Our teams are constantly researching and experimenting with new innovations to deliver maximum performance and efficiency for our users and customers.” Google emphasized its full stack approach and the importance of integrating hardware and software for optimized performance.
AI companies are increasingly developing their own chips to improve model efficiency and address global computing capacity shortages. Efficiency has become crucial amid rising concerns regarding AI-related expenditures impacting the market. Companies are also looking to reduce their dependence on Nvidia, which has historically dominated the AI chip sector.
In June, OpenAI unveiled its first custom chip, named Jalapeño, while Anthropic is reportedly negotiating a chipmaking partnership with Samsung. Investor concerns about Alphabet’s substantial AI investments have increased, with plans to spend between $180 billion and $190 billion this year.
Following the report of the Frozen v2 chip’s enhanced efficiency, Google’s stock rose about 3% in early trading on Monday, easing investor concerns ahead of the company’s upcoming earnings report.
