When Microsoft acquired GitHub for $7.5 billion in June 2018, Satya Nadella framed the deal as a trust transaction. Microsoft would preserve GitHub's independence, keep it open to all clouds and operating systems, and maintain it as a neutral home for global software collaboration.
Eight years later, that promise has collided with the speed of agentic AI. According to Business Insider reporting confirmed by Microsoft, GitHub's infrastructure cannot keep pace with AI-driven coding workflows, forcing Microsoft to route capacity through Amazon Web Services—its largest cloud rival—while it works to migrate the platform to Azure.
The scale of the problem is visible in GitHub's own metrics. Kyle Daigle, GitHub's chief operating officer, reported in April that commits were on pace to hit 14 billion in 2026, up from 1 billion in 2025. That is not just growth; it is a 14-fold increase in a single year.
GitHub's infrastructure team began to feel the pressure sharply in late 2025. In an April availability update, CTO Vlad Fedorov disclosed that GitHub started executing a 10X capacity increase plan in October 2025, only to conclude by February 2026 that the platform needed to design for 30X scale instead. Fedorov tied the acceleration to agentic development workflows that "accelerated sharply in the second half of December 2025."
The migration itself has become a vulnerability. GitHub is trying to shard, migrate, and harden a mature collaboration platform while usage patterns shift beneath it. A May incident exemplified the risk: a schema migration on a heavily used database table cascaded into pull requests, issues, Actions, webhooks, and Git operations. That month alone, GitHub disclosed nine incidents that degraded services.
Microsoft confirmed the broader shift to multi-cloud architecture without naming AWS by name. A spokesperson told Business Insider that "the incredible spike in agentic development" since late 2025 has tested GitHub's limits and that the company is "accelerating the Azure move while exploring a multi-cloud strategy for elasticity and scale." Amazon declined to comment on individual clients.
The decision reflects a hard trade-off. When reliability becomes the constraint, the optics of paying a rival cloud provider rank below the operational risk of GitHub downtime. For developers, downtime is not an abstract cloud capacity problem; it is a blocker to shipping code, reviewing pull requests, and running CI/CD pipelines.
That operational reality has already cost GitHub credibility among high-signal users. Mitchell Hashimoto, co-founder of HashiCorp and creator of Terraform and Vagrant, announced in April that he would move Ghostty, his terminal emulator, off GitHub after 18 years on the platform. His complaint was direct: GitHub was "no longer a place for serious work" when it blocked him for hours per day.
Hashimoto's departure matters because he represents the class of open source maintainer whose projects set the tone for broader developer habits. If those users begin treating GitHub reliability as a tax on their work, competitors do not need to beat GitHub's network outright. They only need to become reliable enough to capture the workflows GitHub fails to keep smooth.
GitHub now faces pressure from AI-native tools like Cursor and Anthropic's Claude Code. An internal Microsoft meeting late last year included discussion of overhauling GitHub to compete with those products. Reliability is no longer a feature; it is a competitive requirement. The AWS pivot is a bet that maintaining uptime matters more than maintaining the unified cloud narrative Microsoft sold in 2018.