
NVIDIA H200 141GB SXM
1.76x the memory of H100 at the same NVLink fabric: H200 SXM stretches context windows and KV-caches without re-architecting the cluster. Ships as modules or pre-integrated HGX trays.
MOQ 1 unit
| Architecture | NVIDIA Hopper (GH100) |
| CUDA cores | 16,896 |
| Memory | 141 GB HBM3e |
| Memory bandwidth | 4.8 TB/s |
| FP8 performance | 3,958 TFLOPS (with sparsity) |
| FP16 performance | 1,979 TFLOPS (with sparsity) |
| FP64 performance | 34 TFLOPS |
| Interconnect | NVLink 4 - 900 GB/s |
| MIG support | Up to 7 instances |
| TDP | Up to 700 W (configurable) |
| Form factor | SXM module for HGX H200 baseboards |
| Warranty | 12 months, SYNDROME AG |
1.76x the memory of H100 at the same NVLink fabric: H200 SXM stretches context windows and KV-caches without re-architecting the cluster. Ships as modules or pre-integrated HGX trays.
This SKU ships against confirmed allocation — the sourcing desk validates availability per order. All prices are USD, FOB our hub of your choice, with tiered discounts from 10 units. Each shipment leaves with serial records and export paperwork prepared by our team.
Air and sea freight from Singapore or Dubai under EXW, FOB, CIF or DDP (Incoterms 2020). Export clearance and shipping documents are handled by our hubs. Commercial terms are structured individually for each deal and jurisdiction — your manager will walk you through the options.
New hardware carries a 12-month warranty (refurbished — 6 months). RMA is handled through the nearest hub with advance replacement from stock where available. Serial numbers are recorded for every unit that leaves the warehouse.