When starting a new project, when wouldn't you want to build in a cloud?

My background: As a newer face when it comes to software development, from the onset AWS was pretty much what I learned and has been my go to since for deployment. I'm certified with them, and feel competent with the basic IT principles the core products are built on. My Question: While cloud computing has obviously created opportunities for products of all sizes, it typically come with trade-offs that are very unique depending on the products composition and demand. What kinds of products or projects have you worked on where cloud computing didn't work out or was unviable?