← Founder Notes
Archive

Gpu clouds just raised prices into the “inference is free” narrative. nebius’s on-demand rates go…

Yethikrishna ROriginal on Threads

gpu clouds just raised prices into the “inference is free” narrative. nebius’s on-demand rates go up 17-21% on oct 1 — h100 to $4.50/hr, b300 to $9.50 — its second hike in three months, with cpu +25%.

model api prices fall every week; the gpus under them don’t.

Context

Verified source: Nebius's price page lists on-demand rates effective Oct 1, 2026, with HGX H100 at $4.50 and HGX B300 at $9.50 per GPU-hour. The Inference reported on Sep 18 that rates rise 17-21% from Oct 1, with H100 up from $3.85 to $4.50. GPU Economy read the same page change on Oct 1 as +16.9%.

Nebius's docs say prices also change for B200, H200 and the vCPU and RAM of a CPU-only platform from Oct 1, and that preemptible pricing changes from Oct 8.

How it compares

The Oct 1 effective date, the 17-21% range, H100 at $4.50 and B300 at $9.50 match Nebius and The Inference; the B300 rate was $7.85 before. That this is a second hike in three months and that CPU rises 25% were not seen in the sources read, so unsupported here, not refuted. The docs confirm CPU-side price changes without the percentage. 'The gpus under them don't' falling is the author's opinion.

Related work

Watch next

  • Read the Nebius docs page for the before and after table across GPU models.

Sources

  1. NVIDIA GPU Pricing, Nebiusnebius.com
  2. Compute pricing in Nebius AI Cloud, Nebius docsdocs.nebius.com
  3. Nebius raises hourly GPU rates up to 21% from 1st October, The Inferencetheinference.org

Provenance

The note above is reproduced unedited from the original post, first published on Threads on 1 October 2026 at 06:37 IST. Sources are the papers and datasets the note draws on.

View the original post
Embed this note
<iframe src="https://founder.myndlabs.tech/notes/embed/gpu-clouds-just-raised-prices-into-the-inference-Dd7sWiZDDCy" width="480" height="420" style="border:0;max-width:100%" loading="lazy" title="Gpu clouds just raised prices into the “inference is free” narrative. nebius’s on-demand rates go…"></iframe>

More notes