Logos identify the maker. Blade AI is an independent supplier.
Live deal · NVIDIA A100
NVIDIA A100 GPU racks, complete systems, available now
Complete ZT Systems Gen 7.0 48U racks, each with four GPU compute servers carrying eight NVIDIA A100 40GB SXM4 GPUs on the HGX A100 board: 32 GPUs and 1.28TB of GPU memory per rack, with utility servers, a 100GbE switch, management network, PDUs and cabling. Available from a single rack, about 1 week after compliance. A proven platform for inference, fine-tuning and HPC on the CUDA software you already run. Complete systems, not loose GPUs.
- 32A100 40GB SXM4 GPUs per rack
- 1.28TBGPU memory per rack (320GB per server)
- 60 racks1,920 GPUs, 76.8TB of GPU memory
- About 1 weekFrom compliance, from a single rack
The lead-time gap
How fast this one ships
-
NVIDIA A100 racks
Complete racks, not loose GPUs
Blade AI: ~1 wkThe market: 16–20 wks~17 weeks sooner
Typical lead time from end-user compliance approval. Confirmed on the quotation.
Specification
| Rack | ZT Systems Gen 7.0, 48U, rails, panels and intra-rack cabling |
|---|---|
| GPU compute servers | 4 x ZT MSF-053000 per rack |
| Utility servers | 3 x ZT MSF-052999 per rack |
| Switching | 1 x Arista DCS-7060CX-32S 100GbE, 1 x management network switch |
| Power | 4 x rack PDUs |
| Per compute server: GPU | 8 x NVIDIA A100 40GB SXM4 on the NVIDIA HGX A100 8-GPU board |
| Per compute server: GPU memory | 320GB HBM2 |
| Per compute server: processors | 2 x AMD EPYC (Rome) 64-core, 1.4GHz |
| Per compute server: memory | 1TB DDR4-2933 (16 x 64GB) |
| Per compute server: network | 8 x NVIDIA ConnectX-6 200GbE |
| Per compute server: power | 6 x 3,000W supplies |
| Lead time | About 1 week after end-user compliance clears |
| Quantity | 60 complete racks (1,920 GPUs), subject to prior sale |
| Price | Confirmed on the quotation |
- Complete racks, not loose GPUs. Available from a single rack.
- Condition, test reports and warranty are confirmed on the quotation.
- Offered subject to prior sale and end-user compliance. Configuration from the rack bill of materials; rack-level components are confirmed in the transaction documents.
How it compares
- Runs the CUDA stack you already use, with no porting.
- NVLink between the eight GPUs in each server on the HGX A100 board.
- 8 x NVIDIA ConnectX-6 200GbE per compute server for scale-out.
Posted 13 November 2026. Specification as supplied. E&OE.
Other live deals
- MiTAC 8-GPU servers with AMD Instinct MI350X, available nowG8825Z5: 8 x AMD Instinct MI350X 288GB, 2 x EPYC 9575F, 3TB DDR5, 8 x 400G, 8U
- Supermicro 8-GPU servers with AMD Instinct MI300X, about 6 to 8 weeksAS-8125GS-TNMR2: 8 x AMD Instinct MI300X 192GB, 2 x EPYC 9654, 2.3TB DDR5, 4 x 400GbE, 8U
- Supermicro 10-GPU servers with AMD Instinct MI350P, available nowAS-5126GS-TNRT2: 10 x AMD Instinct MI350P 144GB, 2 x EPYC 9575F, 2TB DDR5-6400, 5U
Register interest in this deal
NVIDIA A100 GPU racks, complete systems, available now: ZT Systems Gen 7.0 rack. We reply within one working day with the quantity still open, the price and the lead time.
GPU systems are export-controlled. Before we quote we ask who the end user is and where the hardware will be installed. We do not supply the destinations listed on our export compliance page.
How a live deal works
- Register interest with the quantity, the site and the date you need it.
- We reply within one working day with the specification, the quantity still open and the price.
- Mutual NDA, then end-user and site documents for export screening. We may decline a deal.
- Your purchase order and deposit go to the seller of record. Serial numbers and photos reach you before the balance is due.
