Many services to build things on the internet are really cheap to get started. Many even offer a free tiers that are more than enough for small projects to get started with, for hobby projects often more than enough to keep running them indefinitely.
These cheap prices can however be deceiving. Once you start having more data, more traffic and a lot of requests the price can go up really fast. The scale of things on the internet is huge.
What might seem like a waste of time optimizing some resources on a smaller scale might yield huge savings when things get big. This might suggest that it would be a good idea to start the optimization work as soon as possible. But unless you know for sure you'll hit those millions of users it will be a waste of time. Making such optimisations isn't cheap either. It's better to wait until the running costs are becoming big enough that those savings from making things run more efficient will actually make it worth the work.