GitHub Enterprise now includes Copilot, and only AI usage is paid
IA en un minuto newsroom · Editor: Jon Elgezabal
In 30 seconds
With GitHub Enterprise, every licensed user now has Copilot, and the company pays solely for AI consumption, measured against a budget. For now it must be requested from the GitHub or Microsoft sales team for Enterprise Cloud, and GitHub will share self-serve dates within weeks. GitHub keeps the seat even if the budget is used up: code suggestions go on, but chat and review stop.
GitHub has announced that GitHub Enterprise now includes Copilot, its AI coding assistant, for every licensed user. According to its changelog post, published on October 28 during Universe 2026, the company's event, until now enterprises had to watch two things: how many Copilot seats they were paying for and how much AI those users consumed. With Copilot built into the platform, the main thing to manage becomes usage against the enterprise budget.
AI usage is metered from the first request against that budget, using the policies, alerts and attribution the company already has. There is no included allowance to track or top up. Anyone in the account without a GitHub Enterprise license can reach the same thing via Copilot Standalone, priced at $10 per user a month. It offers the same features and models as Copilot Enterprise, minus the monthly AI credits that come with each seat.
For now it is available only for GitHub Enterprise Cloud through the GitHub or Microsoft account team, which can set up a new account on this model or move an existing one. Those who manage their own subscription cannot yet: GitHub will open it in stages and will give dates in the coming weeks, with no change for current subscriptions. Copilot Business and Copilot Enterprise remain on sale.
If the budget runs out, the seat is not lost. Code completion (suggestions as you type) and next edit suggestions (the proposal of the next change in the code) keep working. Chat, code review and the cloud version of Copilot, however, require budget to be available.
In its Universe roundup, GitHub says that so far this year it has put in 4 million CPU cores, 160 PB of fast storage and 50 Tb/s of network capacity, while it moves to a multi-region setup in Azure.
Why it matters · analysis and opinion
The change moves the question from how many licenses to buy to how much AI will be spent, and it forces whoever administers the platform to decide who can consume, with what limit and with what alerts. It is a convenient model for giving Copilot to the whole team, but hard to predict if it is not measured: with no included allowance, every request counts from the first one. It is worth estimating the real consumption of a pilot group and comparing it with what is paid today for Copilot Business or Copilot Enterprise seats, which remain on sale. It is also worth telling the team that, when the budget runs out, chat and code review stop while suggestions continue, so that nobody takes it for a breakdown.
Official source: GitHub · Written with the help of AI: how we make the news
