Google is developing a new server chip, informally named Frozen v2, to integrate Gemini AI models directly into hardware. The Information reports the project aims to resolve internal computing capacity shortages.
The custom hardware could deliver six to 10 times more efficiency than Google’s current offerings. This strategy reduces reliance on third-party chipmakers and lowers AI computation costs.
The news arrives ahead of Alphabet’s quarterly earnings report. Investors are currently monitoring the company’s ability to monetize its massive AI investments.