AI Compute RadarAll GPU classes · GPU prices per hour
CoreWeave GPU prices per hour
What CoreWeave lists per GPU-hour right now for the classes it sells, each with the instance the price sits in, and what the list price did over the last weeks. Read four times a day; described, never forecast.
Radar Rent Index
What CoreWeave prices did
The same rule applied to CoreWeave's own daily series: the current window against the one before it, and against the first week collected. A class with too few days says so instead of showing a change.
| GPU | This week | vs. last week | Since start | Offers | Day band $/h |
|---|---|---|---|---|---|
| L40Slatest day Nov 2 | $2.25median of 7 days | —collecting | —collecting | —no comparison yet | —min–p75 of the latest day |
| A100latest day Nov 2 | $2.70median of 7 days | —collecting | —collecting | —no comparison yet | —min–p75 of the latest day |
| H100latest day Nov 2 | $6.16median of 7 days | —collecting | —collecting | —no comparison yet | —min–p75 of the latest day |
| H200latest day Nov 2 | $6.30median of 7 days | —collecting | —collecting | —no comparison yet | —min–p75 of the latest day |
| B200latest day Nov 2 | $8.60median of 7 days | —collecting | —collecting | —no comparison yet | —min–p75 of the latest day |
No day was excluded. A 30-day change appears once thirty days are on record, from Nov 22.
A day is the median of its hourly medians, counting only hours with at least five offers (verified hosts on Vast.ai, free hosts on Clore.ai). A week is the median of its days and needs at least four of them. Words follow fixed thresholds: up or down from ±8 %, sharply from ±20 %; offers up or down from ±15 %. How the index is made →
Price history
Reading CoreWeave's prices
As of today's readingsOn November 2, 2026 CoreWeave listed 5 of the 8 tracked classes: L40S $2.25, A100 $2.70, H100 $6.16, H200 $6.30, B200 $8.60 per GPU-hour. It was not the cheapest source on any class that day. Its widest distance to the cheapest reading was on the A100: $2.70 against $0.73 at Vast.ai, 3.7 times as much. CoreWeave has been read here since October 24, 2026.
CoreWeave is a GPU cloud built around whole machines: it sells its datacenter GPUs on demand as instances of eight cards, with the CPU, memory and local storage of the node attached, at an hourly price for the whole instance. Its public pricing page needs no account, which is what makes it readable here. The number on this site is that instance price divided by its GPU count, the same arithmetic the AWS and Azure rows use, so the three can be read side by side.
Four times a day the site reads the on-demand tables of CoreWeave's pricing page, one per region, and keeps every row that belongs to a tracked class: the HGX H100, H200 and B200, the L40S and the A100. The plain L40 is a different card and stays out, as do the GB200 and GB300 racks and rows priced on request. The lowest per-GPU figure across the regions is the class's price, with the instance and the region behind it named in the row; spot prices are left out.
CoreWeave's figures sit close to the hyperscalers for the H100 and H200, which is the product: eight-card nodes in its own datacenters with the node's full resources. The smallest unit a renter can take is the whole machine, so the row's instance price is the minimum spend, as with Oracle. For a single card the marketplaces and RunPod remain the comparison, and the difference between them and this row is what a dedicated node with support costs over one host's card.
Price moves at CoreWeave, last 30 days
Any change of the list price between complete days; the big clouds move on the scale of months.
No move on record in the last 30 days.