
NVIDIA H100 80GB SXM5
The full-fat Hopper for training clusters: HBM3 at 3.35 TB/s and 900 GB/s NVLink make SXM5 the right call wherever GPUs talk to each other more than to the CPU. We supply modules individually or as matched 8-GPU sets.
MOQ 1 unit
| Architecture | NVIDIA Hopper (GH100) |
| CUDA cores | 16,896 |
| Tensor cores | 528 (4th gen) |
| Memory | 80 GB HBM3 |
| Memory bandwidth | 3.35 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 | SXM5 module for HGX baseboards |
| Warranty | 12 months, SYNDROME AG |
The full-fat Hopper for training clusters: HBM3 at 3.35 TB/s and 900 GB/s NVLink make SXM5 the right call wherever GPUs talk to each other more than to the CPU. We supply modules individually or as matched 8-GPU sets.
SYNDROME AG holds 3 units of this SKU between the Singapore and Dubai warehouses. Pricing is FOB Singapore / Dubai in USD; volume orders from 10 units receive −5…10%. 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.
