Spot pricing is now open and uncapped
Hi John,
We launched spot pricing for preemptible VMs on October 8. The interest from across the industry was larger than we expected, and one question we kept hearing was how the price actually behaves.
Two weeks ago we made that data public. The pricing API is now open to everyone, and live pricing dashboards are on our website. Anyone can see current prices and recent history for any GPU type and region, even from a phone, without logging into the console.
Spot pricing cap removal
Today we're making a second change: we're removing the price cap. Until now, the spot price always stayed one cent below the on-demand rate. From today it can go above it.
The cap was a guardrail for the launch phase. With spot pricing running steadily, we're ready to let the price move freely and reflect live market conditions on Nebius.
Your controls haven't changed
Set a max price and you'll never be charged above it. Choose follow the spot price and price movements won't stop your workload.
If you run preemptible resources as VMs, Serverless AI, or Managed Kubernetes nodes, this is a good moment to review your settings. A max price is the simplest way to rule out costs you aren't prepared to pay.
And if you need GPU capacity urgently and price isn't your constraint, follow the spot price now opens more room than before. Without the cap, you can hold capacity through demand peaks that would previously have stopped you.
Other spot pricing terms remain unchanged. Read the documentation to learn more:
Questions? Reach out to your Nebius representative or send a request to [email protected]. We're happy to help.
Best,
The Nebius team
Follow us to stay tuned:
© 2026 Nebius B.V., Tripolis Park, Burgerweeshuispad 101, 1076 ER Amsterdam, North Holland, Netherlands.
You are receiving this because you’ve opted-in to Nebius mailing lists.