GitHub Copilot users are burning through a month of credits in a couple of days, and they're not happy about it.
In late April, GitHub announced Copilot would move to usage-based billing, with the new model going live on June 1. The monthly prices didn't change on paper. What changed is what you get for them. Premium model and agent requests now draw from a credit pool tied to how much work you ask the model to do, which means a simple chat question and a multi-hour agentic coding session used to cost you the same thing, and now they very much don't.
GitHub's chief product officer Mario Rodriguez framed the shift as a natural consequence of Copilot growing up. It used to be an in-editor assistant. It's now something closer to an autonomous coworker that can run for hours, and GitHub had been eating the inference bill for that on the old flat plan. Fair enough as a business explanation. The problem is what it actually feels like on the ground.
The GitHub community forum has filled up with paying subscribers who say their Pro+ credits are gone in a day or two of normal work. One long-time subscriber, myalban, wrote that he'd "genuinely advocated for it within my team" and that it pained him to say the new quota system was simply broken. Another user, jtabox, did the math and estimated the effective price hike at somewhere between 220% and 310% depending on the model, then made a sharper point about what the new system pushes you toward:
"The thing is this BS change forces you to use dumber models, which in turn leads to stupid mistakes and the agent trying to figure out what it did wrong, spending a boatload of extra tokens until you fix it manually."
That's the part that stings. To stay under budget, developers pick the cheaper model, which does worse work, which then burns more credits cleaning up after itself. GitHub's CTO Vladimir Fedorov told employees that June was "by far our best month ever" for usage, which makes sense: agents that flail cost more than agents that don't.
The bigger story is that this is happening everywhere, not just at GitHub. Salesforce quietly restructured Agentforce around per-action pricing this year. Most enterprise AI tools are drifting toward some version of the same model. Meanwhile a KPMG survey of over 2,100 senior leaders found that 42% have only partial visibility into what their AI is actually costing them, and only 3% of those without cost visibility are hitting their ROI targets. Companies with real cost controls in place hit ROI five times more often.
That's the awkward math behind the whole AI push right now. Companies budgeted for AI expecting it to pay for itself through the savings automation was supposed to deliver first. The savings have been slower to arrive than expected, and the tools are quietly getting more expensive to run in the meantime. When enterprise IT budgets get reviewed later this year, the gap between "AI will save us money" and "AI is a line item that keeps growing" is going to be the conversation.
For two years the pitch has been that AI would pay its own way, first by replacing tedious work and then by funding the next wave of tools with the savings. What's actually happening is that the tools are the thing getting more expensive, and the savings are hard to find on a spreadsheet. GitHub's revolt is small on its own, but it's a preview of what happens across the industry once the fixed-price training wheels come off and everyone starts getting billed for the agents they were told to build their workflows around. In 2026, the companies that win with AI won't be the ones that adopted it fastest. They'll be the ones that actually know what a month of it costs.
