Google plans new AI server chip ‘Frozen v2’ to boost Gemini efficiency: Report
Google is creating a new server chip, 'Frozen v2', to enhance the efficiency of its AI models for users. It will be 6-10 times more efficient than their current AI chips, addressing limitations in computing capacity. The chip will work alongside their Tensor Processing Units and is planned for deployment by 2028. This is part of the growing demand for AI infrastructure, as analysts predict.