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

NVIDIA RTX PRO 5000 Blackwell

NVIDIA · workstation

Change

0 of 34 fields comparable

⇄ swap sides
SpecCerebras Wafer-Scale Engine 3NVIDIA RTX PRO 5000 Blackwell
AI Tops—2,064
Bandwidth GB S—1,344
Cores900,000—
CS4 Introduced2026-08-18—
CS4 Wse3 Turbo Count3—
Cuda Cores—14,080
DIE Size MM246,225—
Displayport Count—4
Displayport Version—2.1
ECC—true
Fabric Bandwidth Pbit S214—
Form factor—4.4 in H x 10.5 in L, dual slot
FP32 TFLOPS—65
Memory Bandwidth PB S21—
Nvdec Count—3
Nvdec GEN—6
Nvenc Count—3
Nvenc GEN—9
ON Chip Sram GB44—
Pcie GEN—5
Peak AI Petaflops125—
Process NodeTSMC 5nm—
RT Cores—110
RT Cores GEN—4
RT Tflops—196
Slot Width—2
TDP (W)↓ better—300
Tensor Cores—440
Tensor Cores GEN—5
Transistors Trillion4—
VRAM (GB)—48
Vram GB ALT—72
Vram Type—GDDR7 ECC
Wse3 Turbo Peak AI Petaflops250—