Chips · accelerators · Google
Google TPU v5p · TPU
459 teraflops of dense BF16 compute a chip; Google publishes no power figure
Shipped 9 Apr 2024; shipped, no dated promise on record. Epoch AI names it at 10 AI data centers and 1 GPU cluster. Google across the site.
Embed
The frame credits Google, Epoch AI, AI data centers and Epoch AI, Data on GPU clusters and links back to its page. Preview it.
Cite
Each figure in Google’s own words
one chip, dense compute; memory is high-bandwidth memory- BF16, dense
- 459 TFLOPS
Peak compute per chip (BF16) (TFLOPs) 459
TPU v5p - FP8, dense
- 459 TFLOPS
Peak compute per chip (FP8) (TFLOPs) 459
TPU v5p - Memory
- 95 GiB
HBM capacity per chip (GiB) 95
TPU v5p - Memory bandwidth
- 2.77 TB/s
HBM bandwidth per chip (GBps) 2765
TPU v5p (2765 GB/s on the page) - List price
- $4.20 a chip-hour
TPU v5p us-east5 Columbus $4.20
Cloud TPU pricing (on-demand, us-east5 (first US region listed), per chip-hour); 109 TFLOP-hours of dense 16-bit compute a dollar
The v5p page gives the same FP8 figure as BF16 (459). The 2023-12-06 announcement gave no dated availability window, so there is no promise.
The shipment
each date links to the page it is quoted from- 9 Apr 2024 Shipped: Cloud TPU v5p generally available.
We're thrilled to announce the general availability of Cloud TPU v5p
What's new with Google Cloud's AI Hypercomputer architecture
Named at 10 AI data centers
H100-equivalents and power are the whole site’s today, with every chip it holds| Site | Owner | Country | H100e | MW | Chips |
|---|---|---|---|---|---|
| Google Pryor (North) | United States | 636,685 | 442 | TPU v5p, TPU v5e | |
| Google Columbus | United States | 331,986 | 364 | TPU v5e, TPU v5p, TPU v6e, TPU v7 | |
| Google New Albany | United States | 323,901 | 400 | TPU v5e, TPU v5p, TPU v6e | |
| Google Mesa | United States | 289,540 | 237 | TPU v6e, TPU v5p | |
| Google Council Bluffs (East) | United States | 255,685 | 284 | TPU v5e, TPU v6e, TPU v5p, TPU v7 | |
| Google Bristow | United States | 209,861 | 286 | TPU v5e, TPU v5p, TPU v6e, TPU v7 | |
| Google Omaha | United States | 198,585 | 284 | TPU v5e, TPU v5p, TPU v6e, TPU v7 | |
| Google The Dalles | United States | 175,341 | 185 | TPU v6e, TPU v5p, TPU v5e, TPU v7 | |
| Google Storey County | United States | 133,906 | 214 | TPU v5e, TPU v4, TPU v6e, TPU v5p | |
| Google Papillion | United States | 100,556 | 284 | TPU v4, TPU v5e, TPU v5p |
8,960 chips in 1 GPU cluster
Epoch AI’s list of AI supercomputers, which stops in July 2025| Cluster | Owner | Country | Chips | First operational |
|---|---|---|---|---|
| Google Hypercomputer TPU v5p pod | Not given | 8,960 | 2023-12-06 |
Epoch AI’s names for this chip: TPU v5p; Google TPU v5p.
Other Google chips
- TPU v4 275 teraflops of dense BF16 compute a chip; Google publishes no power figure
- TPU v5e 197 teraflops of dense BF16 compute a chip; Google publishes no power figure
- TPU v6e 918 teraflops of dense BF16 compute a chip; Google publishes no power figure
- TPU7x (Ironwood) 2,307 teraflops of dense BF16 compute a chip; Google publishes no power figure
- Google’s own pages, each linked beside the figure or date it gives. Each page was read and each quote checked on 23 Oct 2026.
- Where it runs: Epoch AI, AI data centers and Epoch AI, GPU clusters, CC BY 4.0.