Copilot just 9x'd Sonnet and 27x'd Opus and teams have no idea

GitHub Copilot's recent multiplier table update has sent shockwaves through enterprise teams, revealing a tectonic shift in AI pricing. Opus 4.6's multiplier skyrocketed from 3x to 27x, while Sonnet 4.6 went from 1x to 9x. These aren't arbitrary changes—they represent the true compute cost that Microsoft and Anthropic have been subsidizing for months. The AI companies have been eating the difference between what compute costs and what users pay, but that model has become unsustainable as compute costs for Copilot nearly doubled week-over-week since January.

Starting June 1, GitHub moves to full usage-based billing, and the multiplier hike is just the warning shot. Actual dollar charges will hit corporate cards, traced back to individual usage patterns that nobody thought to govern. Teams that have been treating frontier model access as unlimited, running Opus on everything from code review to boilerplate, will face a rude awakening. Engineering managers will need to explain to finance why the AI budget is 15x over forecast. Every major provider—OpenAI, Anthropic, Cursor—is running the same playbook: the flat-rate era is being unwound in real time, and the new pricing structures are designed to make heavy agentic usage reflect its true cost.