Can we autoscale past 100 GPUs?
Reading the serverless documentation, under the autoscale section, it says "Dynamically scale workers from 0 to 100 on the Secure Cloud platform, which is highly available and distributed globally. This provides users with the computational resources exactly when needed." Not sure if 0 to 100 is meant literally or figuratively-
Our current provider has around 50 H100s available so this is an active point of investigation for us.
TLDR: Can we scale past 100 GPUs on enterprise plans? Is there an enterprise POC I can reach out to?
2 Replies
yes you can scale past 100, we have some users going up to couple 100
0 to 100 don't take it literally, maybe we should change it 😄Thank you @flash-singh, about to run some initial tests shortly and wanted to make sure that wasn't a long-term blocker