Developer Shawn Wang (swyx) calls the 2026 outage problem structural, not capacity, and Dohmke's startup Entire mirrors GitHub repos into regional hubs so coding agents read from nearby infrastructure.
"AI systems interact with code in fundamentally different ways than humans do, and the infrastructure needs to change at a deep level." Developer and infrastructure commentator Shawn Wang, known online as swyx, used that framing to diagnose GitHub's 2026 outage crisis as architectural, not a capacity problem.
GitHub's centralized US origin was built for human developers who push code occasionally and read it on demand. Agentic coding tools generate hundreds of concurrent reads per second across geographies. That traffic profile is structurally incompatible with a single-origin Git host. GitHub's own availability updates, the platform's Copilot sign-up pause, and agentic workflow restrictions document the strain.
Former GitHub CEO Thomas Dohmke's answer is Entire, a preview distributed Git network that mirrors GitHub repositories into regional hubs so coding agents pull from nearby infrastructure while code stays on GitHub. The startup raised a $60M seed at a $300M valuation in February 2026, among the largest dev-tool seed rounds on record, and integrates with Codex, Claude Code, Cursor, and GitHub Copilot. Company benchmarks claim roughly 570,000 shallow clones per hour and 586 pushes per second sustained, both self-reported figures outlined in launch coverage. Entire plans to open-source its backend.
Code delivery is becoming a content-distribution problem. Whether GitHub or a cloud provider builds the mirror layer first is the next signal to watch.