H200 SXM 141GB
Hopper with 141 GB. The inference workhorse of 2026 and the easiest SXM part to actually get.
Allocation. Lead time 3 to 6 weeks.
Volume pricing
| Units | Discount | Unit price |
|---|---|---|
| 1 unit | List | $35,000 |
| 2 to 7 units | 3% off | $33,950 |
| 8 to 31 units | 6% off | $32,900 |
| 32+ units | 10% off | $31,500 |
Unit price at 1: $35,000
Quote only. Price shown is indicative; we confirm price and date in writing before you pay.
Request quote
The full sheet.
Figures from the manufacturer datasheet. Price is an indicative street price, shown as from.
- VRAM
- 141 GB
- Memory
- HBM3e
- Bandwidth
- 4.8 TB/s
- FP8
- 3.96 PFLOPS
- FP4
- Not supported
- Board power
- 700 W
- Form factor
- SXM
- NVLink
- Yes
- Architecture
- Hopper
- GPUs
- 1
- Per GB of VRAM
- $248.23
- Availability
- Allocation, 3 to 6 weeks
- Sale
- Quote only
- SKU
- H200-SXM
What it is for.
Workloads, sizing and what ships in the box.
141 GB of HBM3e, and you can actually get it.
Hopper with 141 GB of HBM3e at 4.8 TB/s and 3.96 PFLOPS of FP8 at 700 W. Drop-in for existing HGX H100 designs, so racks, cooling and software carry over. Single modules or HGX H200 8-GPU sets. Allocation, 3 to 6 weeks. From $35,000, about $248 per GB. Eight of them is 1,128 GB in one box.
Good for
- Llama 3.1 70B at FP8 (84 GB with headroom) with 57 GB left for KV cache.
- Qwen2.5 32B at FP16 (77 GB) on one GPU.
- Llama 3.1 405B at FP8 (486 GB) across 4 of an 8-GPU set.
Fit rule: FP16 weights take 2 bytes per parameter, FP8 1 byte, 4-bit 0.5 bytes, plus 20% headroom for KV cache at short context.
Pairs with
- H200 NVL 141GB, same memory on PCIe for air-cooled boxes.
- H100 SXM 80GB, for spares in mixed Hopper fleets.
HGX H200 8-GPU sets available.
How much VRAM per box.
Card count is capped by slots and by the chassis power budget.
Parts bought alongside.
Same workload, a different trade between VRAM, bandwidth and price.