CoreWeave H100 rental price

CoreWeave lists 2 current on-demand H100 rows. Its lowest is $6.16 per GPU-hour, rank 15 of 19 among providers that set their own price. The median of those 19 is $3.90, so CoreWeave sits 58% above the median.

Lowest listed: $6.16/GPU-hr; 8× HGX configuration
Rank: 15 of 19; by lowest solid price
Against median: +58%; median $3.90
Rows: 2; 2 spot, 0 reserved

Rank is among the 19 providers with a current on-demand price they set; marketplace figures, announced prices, spot and reserved rows are not ranked. The difference to the median uses $3.90, the median of the provider lows. USD per GPU-hour, before tax.

Where CoreWeave sits among the providers

19 providers

18 providers, each at its lowest listed on-demand price, USD per GPU-hour

Closest in price

Every on-demand H100 row from CoreWeave

2 rows

CoreWeave on-demand H100 rows with source and read date
Configuration	Billing	$/GPU-hr	$/hr, config	Region	Read	Source
8× HGXNVIDIA HGX H100	not stated	$6.16	$49.24	North America	2026-11-12	page
8× HGXNVIDIA HGX H100	not stated	$6.16	$49.24	Europe	2026-11-12	page

Current rows first, then rows announced for later. A marketplace row is set by hosts, a from chip gives the start date of an announced price.

Size by size against the market

rank is among providers at the same size

CoreWeave lowest H100 price at each configuration size and its rank at that size
Size	$/GPU-hr	$/hr, config	Rank at this size	Lowest at this size
8×	$6.16	$49.24	7 of 10	Vultr, $2.30

CoreWeave lists one size, 8× with a price it sets, so there is no per-GPU change across sizes to show.

Spot and reserved H100 from CoreWeave

not part of the ranking

Spot (2)

CoreWeave spot H100 rows beside its on-demand price
Configuration	Billing	$/GPU-hr	Its on-demand low	Difference	Read	Source
8× HGXNVIDIA HGX H100	not stated	$2.44	$6.16	-60%	2026-11-12	page
8× HGXNVIDIA HGX H100	not stated	$2.46	$6.16	-60%	2026-11-12	page

Spot is capacity that can be interrupted; reserved is paid for over a term. The difference compares each row with CoreWeave's lowest on-demand price of $6.16.

CoreWeave: common questions

How much does CoreWeave charge for an H100 per hour?

$6.16 per GPU-hour for its lowest listed configuration (8× HGX), read 2026-11-12. For that configuration the hourly total is $49.24.

Where does CoreWeave rank for the H100?

Rank 15 of 19 providers by lowest solid on-demand price. The closest providers by price are Nebius ($4.50), Verda ($5.09), Amazon Web Services ($6.88), and Microsoft Azure ($6.98).

Is CoreWeave above or below the median?

CoreWeave's lowest figure of $6.16 is 58% above the median of $3.90, where the median is taken over each ranked provider's lowest price.

How does CoreWeave bill H100 time?

CoreWeave does not state a billing unit on the pages we read for its H100 rows, so none is assumed.

Does CoreWeave list an 8-GPU H100 server?

Yes: $6.16 per GPU, $49.24 per server-hour.

Does CoreWeave list spot or reserved H100 capacity?

2 spot and 0 reserved rows, shown above with the on-demand low beside each.

What CoreWeave states about itself

States no ingress, egress or transfer fees. source
Lists on-demand and spot prices per instance in North America and Europe. source

Pricing page read: https://www.coreweave.com/pricing

Disclosure

H100 Rate Card is published by the Runcrate team, who also run Runcrate, an AI cloud. Runcrate is listed in the table for the H100, marked "ours, live". It is sorted by price like every other row, with no pinned position. Other providers' prices are read from their own public pricing pages. We have no paid placement.