Google announced two systems for its eighth-generation AI chip lineup: TPU 8tx, designed to protect sensitive data during processing, and TPU 8i Max, designed for memory-intensive tasks. Google did not disclose pricing or availability dates. Its performance claims compare the systems with earlier Google hardware, not Nvidia’s latest products.
Google says TPU 8tx uses hardware-level protection to isolate customer data and model parameters during operation, and can cut latency by up to half compared with TPU 8t. That protection cannot prevent an AI agent from mishandling data it is allowed to access. TPU 8i Max combines two chips and doubles memory throughput compared with TPU 8i, according to Google.
