Different models have different strengths and weaknesses, especially here in the early days when models and their capabilities progress several times per year. The apps, programs, and systems based on models need to know how to exploit their specific strengths and weaknesses. So they are not infinitely interchangeable. Over time some of that differentiation will erode, but it will probably take years.

AWS having customers using its own model probably improves AWS's margins, but having multiple models available (e.g. Anthropic's) improves their ability to capture market share. To date, AWS's efforts (e.g. Q, CodeWhisperer) have not met with universal praise. So for at least for the present, it makes sense to bring customers to AWS to "do AI" whether they're using AWS's models or someone else's.