NVIDIA
B200
SXM
Architecture
Blackwell
Provider Marketplace
All Cloud Providers
Estimates only — rates are collected automatically from public provider pages and may be out of date. Prices vary by region, commitment term, and availability, and typically exclude storage, egress, and tax. Confirm current pricing with the provider before purchasing. Last collected .
Architecture
MicroarchitectureBlackwell
Sparse AccelerationSupported (dense performance is ½ sparse spec shown)
Dynamic PrecisionSupported (FP4)
Connectivity & Scaling
InterconnectNVLink
Generationfifth-generation NVIDIA NVLink
PCIe InterfacePCIe Gen5
TopologyNVIDIA NVSwitch
Max GPUs/Node8
Scale-OutInfiniBand (up to 400Gbps) and Ethernet (up to 400GbE)
Software Ecosystem
CUDA12.1.1
DockerDocker Engine; NVIDIA Container Toolkit; NVIDIA Container Runtime for Docker (nvidia-docker2); native GPU support
Kernel Optimoptimized Linux kernel
Server & Deployment
PreconfiguredNVIDIA DGX BasePOD; NVIDIA DGX SuperPOD
Rack-ScaleNVIDIA DGX SuperPOD (configurable with any NVIDIA DGX system)
Multi-GPU Scalability
Workload Readiness
Market Authority
Community Benchmarks
SemiAnalysis InferenceX benchmarks (Q1 2026)
Enterprise Cases
Examples of Successful Enterprise Deployments
Key Strengths
Limitations
Also in the Lineup
Expert Insight
The B200 represents a strategic leap in AI compute. When comparing cloud providers, consider not just the hourly rate, but also the interconnect bandwidth (InfiniBand/NVLink) and regional availability which can significantly impact total cost of ownership for large-scale training.
Glossary Terms
FP32 TFLOPS
VRAM
TDP
Cores
Information updated daily. Cloud pricing subject to vendor availability.