TradingKey — According to a report from The Information citing two people familiar with the matter, Google (GOOGL) is developing a new server chip capable of directly embedding the underlying architecture of the Gemini large model into the chip hardware to significantly enhance the operational efficiency of AI services for users. Internally codenamed “Frozen v2,” this AI inference chip directly addresses Google’s current severe shortage of AI computing power. Sources revealed that the computing capacity gap has triggered internal resource conflicts and even forced Google Cloud to reject numerous orders from external customers. The research and development team estimates that once the chip is officially deployed, the number of tokens processed per unit of power consumption will be 6 to 10 times that of the latest generation of Google’s existing self-developed AI chips.







