The verdict
AWS takes it, 4–3 on rounds.
On the published figures and the independent review sites, AWS takes 4 of 8 rounds. Treat that as a starting point and check the current figures on both pages, because prices and plans move.
Choose CoreWeave if…
- • You want CoreWeave's edge on what it costs to get your data out and how many places it runs
Choose AWS if…
- • You want AWS's edge on the cheapest h100 and the cheapest h200
Round by round
- Round 1The cheapest H100AWS winsCoreWeave5/10
$6.16
per GPU, worked out from $49.24 an hour for an HGX H100 with eight GPUs
- $6.16, against $5.19
Source: coreweave.com ↗
AWS8/10$6.88
p5.48xlarge $55.04/hr for 8x H100 = $6.88 per GPU-hour, on-demand, US East
- worked out from $41.528 an hour for a p5.48xlarge with eight H100s, US regions
Source: aws.amazon.com ↗
AWS wins this round on what the company itself publishes.
- Round 2The cheapest H200AWS winsCoreWeave5/10
$6.31
per GPU, worked out from $50.44 an hour for an HGX H200 with eight GPUs
- $6.31, against $5.97
Source: coreweave.com ↗
AWS8/10Not stated
Claimed '$47.76/hr = $5.97/GPU' unverifiable; no such price published.
- worked out from $47.76 an hour for a p5e.48xlarge with eight H200s
Source: aws.amazon.com ↗
AWS wins this round on what the company itself publishes.
- Round 3What it costs to get your data outCoreWeave winsCoreWeave8/10
Free
data transfer between CoreWeave and the internet is listed as free
Source: coreweave.com ↗
AWS5/10First 100 GB free, then $0.09 per GB
data transfer out to the internet
- First 100 GB free, then $0.09 per GB, against Free
Source: aws.amazon.com ↗
CoreWeave wins this round on what the company itself publishes.
- Round 4How many places it runsCoreWeave winsCoreWeave8/10
51 data centers
operated across North America and Europe, counted as data centers rather than regions
Source: coreweave.com ↗
CoreWeave wins this round on what the company itself publishes.
- Round 5The cheapest A100TieCoreWeave7/10
$2.70
per GPU, worked out from $21.60 an hour for an eight-A100 instance
Source: coreweave.com ↗
AWS7/10$2.74
worked out from $21.958 an hour for a p4d.24xlarge with eight A100s
Source: instances.vantage.sh ↗
Nothing to separate them here.
- Round 6The cheapest L40SAWS winsCoreWeave5/10
$2.25/GPU-hr ($18/hr for 8)
on-demand GPU instances, North America
- $2.25/GPU-hr ($18/hr for 8), against $1.86/hr (g6e.xlarge)
Source: coreweave.com ↗
AWS wins this round on what the published sources say.
- Round 7What storage costs per monthCoreWeave winsCoreWeave8/10
$0.07/GB-mo (file storage)
on-demand GPU instances, North America
Source: coreweave.com ↗
AWS5/10$0.08/GB-mo (EBS gp3)
EC2 on-demand, US East (N. Virginia), Linux
- $0.08/GB-mo (EBS gp3), against $0.07/GB-mo (file storage)
Source: aws.amazon.com ↗
CoreWeave wins this round on what the published sources say.
- Round 8How much you save by committingAWS winsCoreWeave5/10
Up to 60% off
on-demand GPU instances, North America
- Up to 60% off, against Up to 72% off
Source: coreweave.com ↗
AWS wins this round on what the published sources say.
Side by side
- per GPU, worked out from $21.60 an hour for an eight-A100 instance
- on-demand GPU instances, North America
- AWS Regions
- worked out from $41.528 an hour for a p5.48xlarge with eight H100s, US regions
- worked out from $47.76 an hour for a p5e.48xlarge with eight H200s
- $6.16, against $1.99
- $2.70, against $1.19
- $6.31, against $3.59
- $8.60, against $5.98
- $5.19, against $1.99
- $2.74, against $1.19
- $5.97, against $3.59
- First 100 GB free, then $0.09 per GB, against No egress fees