TPU v5p Pod
TPU v5p Pod is a Google AI system with 8,960 Google TPU v5p and 2,240 Custom host.
Built from BenchGecko data as of April 14, 2026 · updates when the data changes
TPU v5p Pod is a Google AI system with 8,960 Google TPU v5p and 2,240 Custom host.
Basic
Google's high-performance TPU pod for large-scale training. 8,960 TPU v5p chips in a single pod connected via ICI 3.0 fabric. Powers Gemini 1.5 and internal Google AI training. Only available on Google Cloud. The largest single training domain outside of xAI's Colossus.
Deep
The dataset lists 8,100 PFLOPS of FP8 compute and 4,050 PFLOPS at BF16. It has 860 TB of HBM3. Interconnect: ICI 3.0. Power: 2,500 kW, liquid cooled.
Expert
Known deployments: Google DeepMind and Anthropic. Figures come from spec sheets and disclosures in the curated systems dataset, dated on the systems page.
Depending on why you're here
- ·8,960 × Google TPU v5p
- ·Interconnect: ICI 3.0
- ·liquid cooling · 2,500 kW
- ·Compare systems on /systems/google-tpu-v5p-pod
- ·Deployed by Google DeepMind and Anthropic
- ·TPU v5p Pod is a big AI computer built from many Google TPU v5p chips
Frequently Asked Questions
Read the primary sources
- Google Cloud TPU v5p blogcloud.google.com