Azure NCads H100 v5 up for rent

This is what scarcity looks like once it becomes liquid: the quota itself becomes the product. For teams blocked by Azure capacity, the fastest path is now a rented VM, not a portal request.

• –Microsoft's NCads H100 v5 docs position the series for applied AI training and batch inference, which fits the pitch well: https://learn.microsoft.com/en-us/azure/virtual-machines/sizes/gpu-accelerated/ncadsh100v5-series
• –Microsoft lists Standard_NC40ads_H100_v5 at 40 vCPUs, 320 GiB memory, and 1 H100 GPU with 94 GB accelerator memory, so the post's 80GB wording should be verified before paying.
• –Hourly/daily access with root is attractive for short-lived experiments, but trust and tenant isolation become the real buying criteria.
• –This feels more like a capacity-market signal than a startup announcement.
• –Reddit listing: https://www.reddit.com/r/LocalLLaMA/comments/1s387b8/anyone_need_h100_gpu_time_have_spare_azure_quota/