B200 SXM — specs & cloud pricing
The B200 SXM is a datacenter accelerator from NVIDIA. Live specs and modeled cloud $/hr from the Mintok silicon catalog.
Cloud rate
$8per chip-hour (modeled)
Specifications
- Vendor
- NVIDIA
- Class
- Datacenter accelerator
- Architecture
- Blackwell
- BF16 throughput
- 4,500 TFLOPS
- On-package memory
- 192 GB
- Memory bandwidth
- 8,000 GB/s
- TDP
- 1,000 W
More NVIDIA silicon
- GB300 NVL725,000 TFLOPS · 288GB
- Rubin Ultra VR3005,000 TFLOPS · 1,024GB
- Rubin VR2002,500 TFLOPS · 288GB
- NVIDIA B2002,250 TFLOPS · 192GB
- H100 SXM1,979 TFLOPS · 80GB
- H200 SXM1,979 TFLOPS · 141GB
FAQ
How much does the B200 SXM cost per hour in the cloud?
The B200 SXM rents for about $8 per chip-hour on cloud providers (modeled from public rates). Own-vs-rent economics — capex, lease and TCO — are in the Mintok sizing report and rack teardowns.
How much memory does the B200 SXM have?
The B200 SXM has 192 GB of on-package memory at 8,000 GB/s bandwidth.
What is the B200 SXM's BF16 compute throughput?
The B200 SXM delivers 4,500 BF16 TFLOPS at a 1,000 W TDP.
Is the B200 SXM a datacenter or edge accelerator?
The B200 SXM is a datacenter accelerator (1,000 W TDP).
Cost this chip in a fleet
Size a deployment on the B200 SXM, or see the full priced rack BOM it sits in.
Specs and modeled rates maintained by Mintok's catalog watchers from public sources. Not investment advice. Methodology.