GMI Cloud vs Oblivus

GB200 NVL72, GB200 NVL4 and HGX B300 systems available alongside H100/H200

Managed inference platform that runs models on top of the underlying GPU fleet

Both hourly on-demand and longer-term private cloud reservations are published

Per-GPU rates published for each instance type, with multi-GPU pricing scaling linearly

H100 SXM5, NVLink and PCIe SKUs published separately for tradeoffs in interconnect

Significant savings versus on-demand for longer-term commitments