Microsoft turns to Amazon AWS to fix GitHub capacity crisis amid AI surge

GitHub has suffered dozens of major outages in 2026 amid surging AI coding demand

Microsoft turns to Amazon AWS to fix GitHub capacity crisis amid AI surge

Microsoft is quietly turning to its biggest cloud rival, Amazon Web Services, to shore up capacity on its GitHub coding platform following a wave of AI-driven outages in 2026. Two people familiar with the plans told Business Insider that the move was prompted by an unprecedented surge in developer activity fuelled by AI coding tools.

Business Insider first reported the arrangement, which marks a significant strategic shift for the tech giant.

A platform overwhelmed by AI

Microsoft acquired GitHub in 2018, and as an independent company, GitHub had largely run its own data centres. Microsoft had planned to migrate the platform entirely to its Azure cloud service by 2027, but a dramatic rise in AI-assisted software development has complicated that timeline.

GitHub Chief Operating Officer Kyle Daigle wrote on X in April that GitHub commits — records of code changes that serve as a proxy for overall development activity — were on pace to reach 14 billion in 2026, up sharply from 1 billion in 2025. That explosion in activity has strained the platform's compute resources to breaking point.

Microsoft confirms multi-cloud approach

A Microsoft spokesperson confirmed that GitHub is now drawing on multiple cloud providers but declined to comment specifically on any Amazon involvement. "The incredible spike in agentic development that began late last year has tested our infrastructure's limits," the spokesperson said. To address the strain, Microsoft is "both accelerating our move to Azure and continuing to explore a multi-cloud strategy to ensure we have the future capacity, compute elasticity and horizontal scale required to support continued growth," they added.

An Amazon spokesperson said the company does not comment on individual clients, but noted that "customers choose AWS because they need global infrastructure that performs reliably, securely, and efficiently at scale, and we're committed to providing the best performance no matter the workload."

Keeping GitHub running

The urgency behind the decision is clear from GitHub's recent track record. The platform has suffered dozens of major outages in 2026, prompting frustration across the developer community.

Mitchell Hashimoto, Co-Founder of startup HashiCorp, wrote in April that GitHub was "no longer a place for serious work if it just blocks you out for hours per day, every day." The outages have also intensified competitive pressure on the platform, which faces growing threats from AI-powered coding tools including Cursor and Anthropic's Claude Code. In an internal meeting late last year, a Microsoft executive discussed the need to overhaul GitHub to compete with those rivals, according to audio reviewed by Business Insider.

Big tech turns to rivals for capacity

Microsoft's arrangement with Amazon is not an isolated case. AI demand across the industry has grown so intense that major tech companies are increasingly striking capacity deals with competitors.

Earlier this month, SpaceX and Google disclosed a new agreement in which Google will pay SpaceX $920 million a month for AI compute capacity from October 2026 through to June 2029. That deal followed Google's own cloud division agreeing to sell AI compute capacity to Anthropic just two months prior.

Microsoft's wider infrastructure challenge

The reliance on AWS is a notable admission for Microsoft, which is locked in a fierce battle with Amazon for cloud market share. Directing revenue to a direct rival, rather than absorbing the load through Azure, is far from an ideal outcome for the company.

Microsoft recently projected that its capital expenditure for the 2026 calendar year will reach $190 billion, largely to expand data centre capacity. However, many of those data centre projects are facing delays, and Microsoft must simultaneously support a broad range of high-priority AI products and services with its available compute resources.