Skip to content
MCS · Rigs · Compare

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

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

Rig A

Cerebras Wafer-Scale Engine 3

Cerebras · accelerator

Change

Rig B

AWS Trainium2

AWS · cloud-model

Change

0 of 30 fields comparable

⇄ swap sides
SpecCerebras Wafer-Scale Engine 3AWS Trainium2
Bf16 Tflops PER Chip—667
Chips PER Trn2 Instance—16
Chips PER Ultraserver—64
Cores900,000—
CS4 Introduced2026-08-18—
CS4 Wse3 Turbo Count3—
DIE Size MM246,225—
Fabric Bandwidth Pbit S214—
FP8 Dense Petaflops PER Chip—1.3
FP8 Sparse Petaflops PER Chip—5.2
HBM Bandwidth TB S PER Chip—2.9
HBM Capacity GB PER Chip—96
HBM Type—HBM3
Memory Bandwidth PB S21—
Neuroncores PER Chip—8
ON Chip Sram GB44—
Peak AI Petaflops125—
Process NodeTSMC 5nm—
Trainium3 GA—2025-12-02
Trainium3 HBM Bandwidth TB S—4.9
Trainium3 Hbm3e GB PER Chip—144
Transistors Trillion4—
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
Wse3 Turbo Peak AI Petaflops250—