Google launches new AI chips to increase security and computing power

The TPU 8tx aims to bring safer inferences and computational security at the hardware level. The chip is ideal for companies that need to handle and train sensitive data and features custom AMD CPU hosts to ensure information is stored securely.

The TPU 8i Max was created to eliminate possible memory bottlenecks in AI processing and allow large-scale work. The chip supports massive context windows, increases memory bandwidth for data transmission and allows for greater computational density (greater computing power per square meter and per watt).

The new components expand Google’s TPU line: the 8th generation line was announced in April this year with the TPU 8i and TPU 8t models.

Google reports that the increased use of AI agents was one of the factors in the creation of new chips. The leap from human scale to agentic scale requires greater computational power, especially given the fact that a bot can create complex and continuous workflows.

“As computing moves from human-scale to agent-scale, companies are asking for accelerators that can reason about the massive context windows of AI models. Additionally, some highly regulated customers are asking for specialized encryption capabilities for unique workloads,” explains Google’s Chief Technology Officer and Senior Vice President of AI and Infrastructure, Amin Vahdat.

Acronym for Tensor Processing Unit, TPU is a chip developed by Google for use in machine learning and AI infrastructure.

The component is used in data centers and allows greater computing power for inference when training models.

Affiliate Disclosure: Articles published by Silicon Dispatch may include verified affiliate partner links. If you purchase through these links, we may earn an affiliate commission at no additional cost to you.