This is a dedicated watch page for a single video.
A global bank decides to run its new generative AI service exclusively on its private, on-premises servers due to data residency regulations. How does this decision at the Infrastructure layer most significantly constrain its choices at the Model layer?