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

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

Rig A

AWS · cloud-model

Change

Rig B

NVIDIA · accelerator

Change

0 of 33 fields comparable

Spec	AWS Trainium2	NVIDIA B200
Bf16 Tflops PER Chip	667	—
Chip TO Chip Interconnect TB S	—	10
Chips PER Trn2 Instance	16	—
Chips PER Ultraserver	64	—
Fp16 Bf16 Pflops	—	4.5
FP32 TFLOPS	—	75
FP4 Pflops Dense	—	9
FP4 Pflops Sparse	—	18
FP64 TFLOPS	—	37
FP8 Dense Petaflops PER Chip	1.3	—
FP8 Pflops	—	9
FP8 Sparse Petaflops PER Chip	5.2	—
HBM (GB)	—	180
HBM Bandwidth TB S PER Chip	2.9	—
HBM Capacity GB PER Chip	96	—
HBM Type	HBM3	—
Int8 Pops	—	9
Memory bandwidth (TB/s)	—	8
Neuroncores PER Chip	8	—
Nvlink Generation	—	5
Nvlink GPU TO GPU TB S	—	1.8
Process Node	—	TSMC 4NP
Trainium3 GA	2025-12-02	—
Trainium3 HBM Bandwidth TB S	4.9	—
Trainium3 Hbm3e GB PER Chip	144	—
Transistor Count Billion	—	208
Trn2 Instance FP8 Petaflops	20.8	—
Trn2 Instance HBM TB	1.5	—
Trn3 Ultraserver FP8 Pflops	362	—
Trn3 Ultraserver MAX Chips	144	—
Ultraserver FP8 Petaflops	83.2	—
Ultraserver Total HBM Bandwidth TB S	185	—
Ultraserver Total HBM TB	6	—