AI hardware
Google TPU v4i
FP16 throughput
1.4×10¹⁴ FLOP/s
Memory
7.5 GB
TDP
175 W
Google TPU v4i is an AI accelerator designed by Google (TPU), released in January 2020. It is built on a 7 nm process, packing 16B transistors.
It delivers 1.4×10¹⁴ FLOP/s of tensor FP16/BF16 throughput. It carries 7.5 GB of memory at 572 GB/s of bandwidth. Its thermal design power is 175 W.
Full record
- Manufacturer
- Type
- TPU
- Release date
- Jan 1, 2020
- Tensor FP16/BF16
- 1.4×10¹⁴ FLOP/s
- Memory
- 7.5 GB
- Memory bandwidth
- 572 GB
- TDP
- 175 W
- Process node
- 7 nm
- Transistors (millions)
- 16,000
- Die size (mm2)
- 400
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