AI hardware
Google TPU v4
FP16 throughput
2.8×10¹⁴ FLOP/s
Memory
29.8 GB
TDP
340 W
Google TPU v4 is an AI accelerator designed by Google (TPU), released in May 2021. It is built on a 7 nm process.
It delivers 2.8×10¹⁴ FLOP/s of tensor FP16/BF16 throughput. It carries 29.8 GB of memory at 1.1 TB/s of bandwidth. Its thermal design power is 340 W.
Full record
- Manufacturer
- Type
- TPU
- Release date
- May 20, 2021
- Tensor FP16/BF16
- 2.8×10¹⁴ FLOP/s
- INT8
- 2.8×10¹⁴ OP/s
- Memory
- 29.8 GB
- Memory bandwidth
- 1.1 TB
- TDP
- 340 W
- Process node
- 7 nm
More from Google
SourceEpoch AI, 'Machine Learning Hardware'. Published online at epoch.ai. Retrieved 2026-07-29 from https://epoch.ai/data/machine-learning-hardware. Licensed under CC BY 4.0.
← All ai hardware