Microsoft Turns to AWS for GitHub Capacity as AI Demand Strains Azure

Microsoft is turning to rival Amazon Web Services to prop up GitHub's cloud capacity as surging AI workloads overwhelm Azure. The move highlights how even the largest hyperscalers are struggling to keep pace with demand driven by agentic development and AI-assisted coding workflows.

GitHub confirmed to The Register that it is "both accelerating our move to Azure and continuing to explore a multi-cloud strategy" to meet demand from a spike in agentic development that began late last year. The coding platform, acquired by Microsoft in 2018, has suffered ongoing reliability issues as AI workflows multiplied.

GitHub's capacity crunch forces multicloud pivot

The decision to supplement Azure with AWS resources is a significant reversal for a company that owns its own cloud infrastructure. Sources told TechRadar Pro that the arrangement involves using AWS compute and storage to handle overflow from GitHub's services. It is unclear whether this is a temporary fix or a permanent multicloud architecture.

• GitHub has faced repeated outages and slowdowns as AI-assisted coding volumes surged.
• The company previously attempted to shift all workloads to Azure but found the cloud couldn't scale fast enough.
• Microsoft's Azure has itself experienced capacity problems in recent months, according to reports.
• GitHub's community growth rate is described by the company as unprecedented.

Shareholder lawsuit adds legal pressure

Separately, Microsoft faces a class action lawsuit filed by the City of St. Clair Shores Police and Fire Retirement System in Seattle federal court. The complaint alleges CEO Satya Nadella and other executives made "materially false and/or misleading" statements about Copilot adoption and AI model competitiveness.

The lawsuit points to Microsoft's fiscal second quarter results, which revealed paid Microsoft 365 Copilot seats had reached only 15 million out of 450 million total users, contradicting earlier bullish statements. Shares dropped more than $48 per share, about 10 percent, after the disclosure. A Microsoft spokesperson told The Register the company believes the claims are without merit and will defend itself vigorously.

Efficiency demands reshape data center planning

These capacity pressures come as the broader industry grapples with power availability and cooling constraints. HPE's Andrew Desrochers recently told Data Center Knowledge that power availability, utility lead times, and water use are now core IT planning priorities, not just facilities concerns. Equipment vendors like Kerun are launching integrated transformer and substation solutions specifically for high-heat AI data center deployments.

Bitdeer is also planning a 750MW data center campus in Ohio but faces a local moratorium on new construction. The confluence of hyperscaler capacity limits, legal scrutiny over AI claims, and physical infrastructure bottlenecks suggests the next phase of AI growth will be constrained not by software innovation but by concrete, copper, and kilowatts.