More modern GPU options

My application is deployed in ORD doing inference on the A100’s, and I would really like to have 80Gb+ options as well as Blackwell based cards I could use for inference.

Right now Fly seems to be falling behind in terms of which GPU’s its offering, and as larger local models become available, I don’t want to end up in a world where my GPU workload is on one providers and my webhosts are on another.

1 Like

we’re not going to add more gpu options, unfortunately: We Were Wrong About GPUs · The Fly Blog

Just for discussion @lillian: do you think Fly might ever drop GPUs?

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.