AI Workloads on Cloud

The debate around cloud repatriation gained traction after Barclays' CIO survey highlighted rising cloud costs as a key concern.

Tom Nolle suggests that effective workload planning can prevent the need for repatriation, potentially saving significant costs. He identifies three main drivers of high cloud expenses: 

1-    Using inappropriate cloud services, 

2-    Overestimating application load leading to excess scalability, and 

3-    Developers’ lack of cost-saving insights when designing applications.

I’ve also emphasized this in a recent LinkedIn post, noting the importance of careful planning for compute-intensive AI workloads and the emerging role of a Cloud Cost Controller in 2025.