Side-by-side, with the numbers that matter.

Two rigs, one table. Higher numbers win in cobalt; TDP and price flip.

Rig A

Google · cloud-model

Change

Rig B

AMD · accelerator

Change

0 of 37 fields comparable

Spec	Google Cloud TPU v5p	AMD Instinct MI355X
Bf16 Matrix Pflops Dense	—	2.5
Bf16 Matrix Pflops Sparse	—	5
Bf16 Tflops PER Chip	459	—
Chips PER POD	8,960	—
Compute Units	—	256
Cooling	—	direct liquid cooling (air-cooled option available)
Form factor	—	OAM
Fp16 Matrix Pflops Dense	—	2.5
Fp16 Matrix Pflops Sparse	—	5
Fp32 Vector Tflops	—	157.3
Fp64 Matrix Tflops	—	78.6
FP8 Matrix Pflops Dense	—	5
FP8 Matrix Pflops Sparse	—	10.1
FP8 Tflops PER Chip	459	—
HBM (GB)	—	288
HBM Bandwidth GB S	2,765	—
HBM Capacity GIB	95	—
HBM Type	—	HBM3E
Helios Gpus PER Rack	—	72
ICI Bandwidth GB S	1,200	—
Memory bandwidth (TB/s)	—	8
Mxfp4 Matrix Pflops	—	10.1
Mxfp6 Matrix Pflops	—	10.1
Pcie GEN	—	5
Process Node NM	—	3
Sparse Cores PER Chip	4	—
Status	available (earlier generation)	—
Successor Cdna Version	—	5
Successor Family	—	AMD Instinct MI400 Series (MI455X)
Successor Launch	—	2026-07-23
Successor Mi455x Bandwidth TB S	—	23.3
Successor Mi455x Hbm4 GB	—	432
Successor Tpu7x Chips PER POD	9,216	—
Successor Tpu7x GA	2026-03-31	—
TBP Watts	—	1,400
Tensor Cores PER Chip	2	—
Topology	3D torus	—