GPU availability & instance types
Are B300/B200 instances bare metal?
They're virtual machines designed to deliver near-bare-metal performance , with full PCIe GPU passthrough that gives you real root and driver control...
What's the difference between on-demand and spot?
On-demand gives you dedicated capacity that stays yours until you delete it. Spot runs on spare capacity at a lower rate but can be reclaimed by Ve...
What does a specific GPU cost?
On-demand and spot rates for every GPU are on the pricing page and in the console when you create an instance. Rates vary by model, configuration, a...
What contract options are available?
Pay as you go - no commitment, billed in 10-minute increments, unused time refunded. Long-term contracts - paid up front at a discount that grows wi...
What is a spot instance, and how much notice do I get before it's reclaimed?
A spot instance runs on spare capacity at a reduced rate, and Verda can reclaim it at any point with no advance notice - the timing can't be predict...
The GPU I want isn't available - when will it come back?
High-demand GPUs sell out quickly and we can't reliably predict when capacity frees up. On the deployment page, click Notify me on the configuration...