Proven Performer is moving serious AI compute: 70 ASUS ESC N8-E11 servers carrying eight NVIDIA H100 80GB GPUs each, 173 NVIDIA Tesla T40 inference accelerators, and 320 brand-new HPE InfiniBand NDR 400Gb/s adapters. Tested, warrantied, and available now — not on a lead time. Private cash/wire buyers only.
| Product | Part / Model | Condition | Warranty | Qty | Location |
|---|---|---|---|---|---|
| ASUS ESC N8-E11 — 8× NVIDIA H100 80GB SXM AI server | ESC N8-E11 | Refurbished | 1 year | 70 | North America |
| NVIDIA Tesla T40 24GB accelerator | 699-2G150-0210-505R | Used, tested | 90 days | 173 | EU |
| HPE InfiniBand NDR 400Gb/s 1-port OSFP PCIe 5.0 x16 | MCX75310AAS · P45641-B21/H21 | New | — | 320 | New Jersey (EXW) |
High-density AI/HPC infrastructure for enterprise and hyperscale deployment — 70 refurbished units with a one-year warranty. Each server: 8× NVIDIA H100 80GB SXM GPUs, dual Intel Xeon Platinum 8468, 2TB DDR5 ECC RDIMM, 2× 960GB + 2× 3.84TB NVMe PCIe 4.0 SSDs, 10Gb LAN, 4+2 3000W 80 PLUS Titanium power.
The GPU tray: eight NVIDIA H100 80GB SXM modules per chassis, NVLink-connected for training and large-model inference. 70 servers means 560 H100 GPUs in one transaction — one of the largest surplus H100 blocks on the market.
173 enterprise inference accelerators — 24GB GDDR6, passive server-ready design, ideal for AI inference, VDI, and media workloads. Used, tested and verified, 90-day warranty, located in the EU. Bulk lot available now.
The fabric side of the AI build-out: 320 genuine HPE InfiniBand NDR adapters — 400Gb/s, single-port OSFP, PCIe 5.0 x16. New in factory packaging, real product photos available, EXW New Jersey. Built for demanding AI/HPC clusters and high-throughput data center environments.
Tell us which line and quantity you're after. Serious inquiries get real pricing the same day — usually within the hour.
Also in stock: data center power & switchgear → · CAT 3516E generators → · everything we're moving →