Google announced two new TPU chips: TPU 8i optimized for AI agent inference and reasoning, and TPU 8t optimized for training large models. Both are designed to enable responsive, multi-step autonomous AI agents. This represents Google's hardware play to support agentic AI at scale.
Infrastructure
We're launching two specialized TPUs for the agentic era.
Google launches TPU 8i and TPU 8t chips purpose-built for agentic AI—inference and training respectively—signaling that specialized silicon will be critical infrastructure for autonomous agent workloads.
Wednesday, April 22, 2026 12:00 PM UTC2 MIN READSOURCE: Google AI BlogBY sys://pipeline
Tags
infrastructure
/// RELATED
InfrastructureApr 22
The eighth-generation TPU: An architecture deep dive
Google's TPU 8t and 8i variants eliminate data-preparation bottlenecks with custom Axion CPUs, delivering specialized training and inference hardware optimized for world models and agentic AI at scale.
ProductsApr 22
Google claims to have all the answers for enterprise AI agent sprawl
Google rebrands Vertex AI as Gemini Enterprise Agent Platform with unified governance, Agent Studio, and memory management to centrally control multi-agent sprawl at enterprise scale.