H200 NVL inference nodes from local stock, DGX Spark available to order, B300-class on honestly-quoted allocation — delivered DDP across the UAE and GCC in days.
Most GPU quotes in the region bounce through global distribution and land at 8–16 week lead times. Haink holds stock in Dubai's free zone and ships H200-class hardware while others are still confirming allocation.
| Platform | Spec | Availability |
|---|---|---|
| NVIDIA H200 NVL | 141 GB HBM3e, passive PCIe, NVLink bridge | To order — screened destinations |
| GPU server nodes | Dell R760xa / HPE DL380a + 2–4× H200 NVL | Base units often from stock |
| NVIDIA DGX Spark | GB10, 128 GB unified — desktop AI | Typically from stock |
| HGX B200 / B300 | 8-GPU training systems | On allocation, quoted realistically |
| L40S / RTX 6000 Ada | Inference and workstation GPUs | Frequently from stock |
| AI fabric | 400G QSFP-DD optics, ConnectX, InfiniBand | Optics same-day capable |
Stock rotates daily — positions are "typically available" and confirmed per request, usually within one business day. Stock guides →
The H200 NVL (passive PCIe, 141 GB) is available to order via our Dubai hub for permitted destinations, subject to end-user screening — it avoids HGX allocation queues entirely. Serial numbers are provided before payment for verification.
An H200 NVL card runs ~$31–32k; a configured dual-GPU inference node from ~$78,000 delivered DDP. B300-class systems are quoted per allocation.
Yes — 2–5 business days typical, with export screening, certificates of origin and destination customs handled.
You can — and wait 8–16 weeks for allocation. Local stock plus regional logistics is the entire point: days instead of quarters, with the same OEM warranties.
Pricing, availability and delivered lead time within one business day.