Google plans new AI server chip ‘Frozen v2’ to boost Gemini efficiency: Report
Google plans to develop a new AI server chip, codenamed 'Frozen v2', to enhance the efficiency of its Gemini AI models by embedding parts of the model directly into hardware. This chip, expected to be six to 10 times more efficient than current models, aims to address limitations in AI computing capacity that have hindered Google Cloud's external partnerships. Deployment is anticipated by 2028, as engineers finalize its design and capabilities.