Vast.ai GPU marketplace (supply-and-demand, interruptible-first)
Fields
supply/vast-ai-marketplaceNotes
A structural record, and the cheapest credible ROUTE to consumer and workstation NVIDIA silicon - the right provider for an RTX 4090/5090 hour or an L40S, and the wrong one for an H100 with guaranteed capacity. Three structural facts from the page: 68+ GPU types, per-second billing with 'no minimum hours, no rounding up', and three tiers - On-Demand (guaranteed uptime, no interruptions, per-second), Interruptible ('Best Value', '50%+ cheaper', 'preemptible - may be reclaimed', for batch training), and Reserved (1/3/6-month terms, guaranteed capacity, up to 50% off, volume discounts, contact sales). The trust tradeoff is the whole story: the hosts are third parties, the supply varies minute to minute, and the page sells on the word 'transparency'. Do not confuse the 'Up to 50% Off' reserved ceiling with the '50%+ cheaper' interruptible tier - they are different mechanisms. If you need an interruptible price for H100 and Vast's grid is reachable to you, this is likely the floor of the entire market.
EXPORT CONTROL (2026-10-04, mechanism-level; sources spl-bis-2026-05-d5-guidance, spl-bis-2026-01-advanced-computing-case-by-case): renting capacity does NOT remove the export-control question, it relocates it. Providing remote access to controlled computing capability can constitute a transfer under the EAR, so the provider must screen the end user; and separately, under the 91 FR 1684 licence regime the consignee must run KYC procedures to prevent unauthorized remote access and must manage and limit IaaS access to the hardware - obligations pushed onto the customer, who must be told to implement them. For a customer headquartered or parented in Country Group D:5 or Macau the 742.6(a)(6)(iii)(A) end-user/parent test can bite even when the capacity is consumed in a permitted region. Recorded true because the underlying hardware is controlled and the provider carries the screening duty, not because access to a US region is itself restricted.
LEAD TIME: null, and correct as null. For a listed on-demand instance type the provisioning delay is effectively zero - capacity is allocated per request - so a week count would be meaningless; the real gate for scarce shapes is either a quota increase or a capacity reservation, not a lead time. Where a published commitment horizon exists it IS recorded: AWS documents that a Capacity Block may be reserved with a start time up to EIGHT WEEKS in the future (source spl-aws-capacity-block-booking-horizon), which is a floor on how far ahead scarce capacity must be committed, not a wait. On-demand and Spot tiers publish no lead time at all; their constraint is availability, and the repo records that as availability=in-stock or a spot/reality note rather than as a fabricated week figure. Probe log in source spl-negative-no-oem-publishes-lead-time.