An eight-GPU NVIDIA HGX H200 server from Supermicro costs in the low-to-mid six figures in pounds sterling excluding VAT, and SuperServer.ai delivers it to the UK in 4 to 5 weeks from order.
GPU allocation changes month to month and the GPU board is most of the cost, so suppliers quote rather than list. A quote fixes the price and lead time in writing for a set period.
In order: the GPU board (H100 80GB or H200 141GB), system memory (1TB is the usual floor, 2TB for heavy CPU preprocessing), networking (eight 400G ConnectX-7 cards for clusters, two 200G for standalone nodes), NVMe storage, and the support term.
SuperServer.ai's target is 4 to 5 weeks from order for single nodes. The industry norm is two months or more. Lead time is confirmed in writing before you commit.
Below about 30% utilisation, yes. Above 60% utilisation for 18 months or more, owning usually wins. An 8× H200 node at current pricing pays back against on-demand cloud in roughly 12 to 16 months before power and colocation.
H200 for inference and for any model above 70B parameters, because of its 141GB of memory. H100 only if the price gap is large and the workload is compute-bound training of smaller models.
Tell us what you are running and we reply with a written quote and lead time. No sales call unless you want one.
Figures on this page are indicative, exclude VAT, and were checked on 1 October 2026. Specifications come from the manufacturers' published data. SuperServer.ai is independent of Supermicro and NVIDIA.