Latest / The Cloud Business Podcast with Fexingo: AWS, Azure, GCP, and Enterprise Infrastructure
Why Cloud Contracts Now Charge for AI Training Job Interruption Costs
Episode 133 of The Cloud Business Podcast dives into a new cloud billing line item: fees for interrupted AI training jobs on spot GPU instances. Lucas and Luna break down how AWS and Azure now charge for each restart, with real numbers from a startup that saw $15,000 in extra costs over a three-month training run. They explain why these fees exist, how to minimize them, and what it means for the economics of large-scale model training. Plus, the hosts discuss strategies like checkpointing and using on-demand instances for critical phases. If you're building AI infrastructure, this…
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.