With AWS supporting up to 10240MB for lambdas and some functions for tasks like PDF generation or data processing like large kmeans calculations with vector embeddings it would be useful to have certain functions at 4, 6, or even 10gb of ram as opposed to the current 2gb cap. I’ve had to splinter projects over to AWS + Vercel working with tools like pulumi or TF to keep things managed and it’s a huge hassle that could be avoided if the max function sizes for the Vercel lambda wrappers were increased to match AWS.
As a plus to whoever on the sales or finance team it would make Vercel some more money on function usage and pull in some ML ops stuff which is at least in my case forced over to AWS.
See Lambda quotas - AWS Lambda for docs on current lambda size range, 128 - 10240 mb as of most recent check