You're overpaying for GPUs if you're running on AWS.
Book a call and we'll give you 100 H100 hours for free. Here's how we can afford to do that without going bankrupt:
On AWS, a 3-year reserved instance for a single H100 GPU costs about $5.40 per hour.
For a 3-year reserved p5.48xlarge instance (8 H100 GPUs)? ~$1.13 million.
On RunPod, a single H100 on-demand, available right now, is just $2.99/hr.
A 6 month 8xH100 pod reservation is just $2.49/hr. Just ~$522k for 3 years. Insane.
100 H100 hours on AWS is $540. We cost less than half - just $250.
Talk to us to learn more 👇
