NVIDIA H200 vs B200
Compare published specifications with current rental availability.
Differences are derived only from verified memory, bandwidth, power, form factor, precision support, software ecosystem and real inventory.
| Field | NVIDIA H200 SXM 141GB | NVIDIA B200 SXM 180GB GPU |
|---|---|---|
| Vendor | NVIDIA | NVIDIA |
| Exact SKU | H200 SXM | B200 SXM |
| Architecture | Hopper | Blackwell |
| Generation | Hopper | Blackwell |
| Category | Data center | Data center |
| Lifecycle | Current | Current |
| VRAM | 141 GB | 180 GB |
| Memory technology | HBM3e | HBM3e |
| Bandwidth | 4,800 GB/s | 8,000 GB/s |
| Power | 700 W | 1000 W |
| Interface | PCIe Gen5 host interface | HGX platform |
| Form factor | SXM · Module | SXM6 · Module |
| Interconnect | NVLink | NVLink 5 |
| MIG / partitioning | Supported | Supported |
| Numeric precision formats | FP64, FP32, TF32, BF16, FP16, FP8, INT8 | FP64, FP32, TF32, BF16, FP16, FP8, FP6, FP4, INT8 |
| Runtime ecosystem | cuda, vllm, sglang, tensorrt-llm | cuda, vllm, sglang, tensorrt-llm |
| Rental availability | Limited capacity | Reference only |
| Monthly rental price | USD 2,345 | Reference only |
Verified differences
- NVIDIA B200 SXM 180GB GPU has the most VRAM in this selection.
- NVIDIA B200 SXM 180GB GPU has the highest recorded memory bandwidth.
- NVIDIA H200 SXM 141GB has the lowest published power value.
Architecture differences
Source-verified facts: NVIDIA H200 SXM 141GB: Hopper / Hopper / SXM · Module · NVIDIA B200 SXM 180GB GPU: Blackwell / Blackwell / SXM6 · Module
VRAM and memory requirements
Source-verified facts: NVIDIA H200 SXM 141GB: 141 GB / HBM3e / 4,800 GB/s · NVIDIA B200 SXM 180GB GPU: 180 GB / HBM3e / 8,000 GB/s
Power and cooling
Source-verified facts: NVIDIA H200 SXM 141GB: 700 W / SXM · Module · NVIDIA B200 SXM 180GB GPU: 1000 W / SXM6 · Module
AI precision support
Source-verified facts: NVIDIA H200 SXM 141GB: FP64, FP32, TF32, BF16, FP16, FP8, INT8 · NVIDIA B200 SXM 180GB GPU: FP64, FP32, TF32, BF16, FP16, FP8, FP6, FP4, INT8
Interface and interconnect
Source-verified facts: NVIDIA H200 SXM 141GB: PCIe Gen5 host interface / NVLink · NVIDIA B200 SXM 180GB GPU: HGX platform / NVLink 5
Multi-GPU considerations
Large variants may require tensor parallelism or multiple nodes. Validate total resident weights, runtime support, interconnect, topology and headroom before choosing a GPU count.
Model and workload fit
Model fit depends on exact checkpoint precision, runtime support, memory headroom and topology. Reference compatibility does not imply FusionPointAI rental inventory.
Rental availability
Source-verified facts: NVIDIA H200 SXM 141GB: Limited capacity / USD 2,345 / month · NVIDIA B200 SXM 180GB GPU: Reference only
Official sources
Last verified: 2026-08-25