Azure Customers up in Arms over ‘Full’ UK South Region

Why It Matters

The shortage hampers UK enterprises that rely on low‑latency, UK‑only cloud services, potentially driving them to competitors or costly multi‑region architectures. It also highlights the trade‑off between AI‑driven growth and core cloud reliability.

Key Takeaways

• Azure UK South VM capacity scarce, especially AMD and GPU
• Copilot AI rollout consuming datacenter power, limiting traditional workloads
• Customers forced to migrate to UK West or delay projects
• Microsoft promises monitoring, but capacity additions not due until 2026
• Future data centre builds in High Wycombe and Harlow may relieve pressure

Pulse Analysis

Azure’s UK South region, the primary hub for many British enterprises, is now a bottleneck for critical workloads. Reports from Reddit and industry forums show customers repeatedly hitting quota walls for AMD‑based VMs, high‑performance compute, and GPU instances—resources essential for everything from virtual desktops to AI model training. The problem is not isolated; one availability zone appears especially starved, forcing firms to either pause migrations or reroute traffic to the single‑datacenter UK West region, which lacks the same resilience and GPU variety.

The root cause appears tied to Microsoft’s strategic push for Copilot AI, which relies on power‑hungry GPU clusters. By prioritising AI‑centric hardware, Microsoft has inadvertently squeezed out traditional cloud services that generate the bulk of its revenue. Analysts note that the rapid deployment of hundreds of GPUs can strain power, cooling, and floor space, turning the region’s capacity into a zero‑sum game. This shift underscores a broader industry tension: balancing the lucrative AI market against the need to maintain dependable infrastructure for existing customers.

Looking ahead, the capacity gap won’t close until at least 2026, when roughly 121 MW of new datacenter power is slated for completion across High Wycombe, Harlow, and Newport sites. In the interim, enterprises may need to diversify across regions, adopt hybrid‑cloud strategies, or negotiate dedicated capacity with Microsoft. The situation serves as a cautionary tale for cloud providers—overcommitting to emerging AI services can erode the reliability that underpins enterprise trust. Companies watching the UK market should monitor Microsoft’s allocation policies and consider contingency plans to avoid operational disruptions.