Vladimir Fedorov says June was GitHub Copilot’s best month ever after billing change
Usage jumped after GitHub switched Copilot from flat per-user pricing to billing based on how much developers use it.

Microsoft-owned GitHub CTO Vladimir Fedorov told employees that June was “by far our best month ever” as Copilot usage surged. For decision-makers, the move signals how pricing mechanics are quickly becoming a competitive weapon in AI coding.
Microsoft-owned GitHub CTO Vladimir Fedorov told employees that June was “by far our best month ever,” and pinned the surge to one decision: how GitHub bills for its AI coding tool, Copilot. In a Wednesday meeting, he said usage jumped significantly after GitHub changed its billing model, while declining to provide the actual numbers because the company’s financial quarter is coming to a close.
The timing matters because GitHub made the switch on June 1. Instead of charging Copilot as a flat rate per user for a fixed number of requests, GitHub moved to billing based on consumption, meaning how much teams actually use the tool. That is the whole story in one sentence, and it explains why executives are paying attention now: in AI coding, “what you pay” and “how much you get” can drive adoption faster than marketing ever will.
This billing change also tees up a broader competition that is getting louder by the week. GitHub’s surge is described as a bright spot for the developer platform as it competes with fast-growing rivals such as Cursor, OpenAI’s Codex, and Anthropic’s Claude Code. These tools all help write, edit, and fix software. In other words, GitHub is not just competing on where engineers store code and collaborate. It is competing on the day-to-day workflow inside the editor, where AI assistance can make teams faster and, crucially, can reduce friction in repetitive coding tasks.
For boards and senior leaders, the second-order implication is simple: in a usage-based pricing world, product performance becomes finance performance. If usage rises after a pricing model shift, procurement behavior changes too. Teams that previously hesitated to expand AI usage because they felt capped by a quota may now feel freer to “turn it up” because billing tracks actual consumption. That can increase both seat expansion and tool intensity, and it can also change how CFOs forecast costs, since expenses become variable rather than fixed. Fedorov also said he personally doesn’t think GitHub needs to raise prices much based on this usage spike, though he did not disclose any definitive plans on pricing.
GitHub’s incentives are also shaped by competitive pressure. GitHub had an early lead, helped by its popularity as a place for engineers to store and manage code and collaborate on projects. But lately, it has faced more competition from Cursor and Anthropic’s Claude Code. The underlying pressure is that switching costs in developer tools are often less about migrating entire repositories and more about migrating workflows. If an engineer’s editor AI feels better, faster, or cheaper in practice, tool adoption can happen even if the code hosting platform remains the same.
Microsoft acquired GitHub in 2018, so Microsoft’s strategy is deeply tied to GitHub as an ecosystem. But Business Insider’s reporting adds another layer to why this month matters beyond growth. Due to increased usage, GitHub experienced dozens of major outages in 2026. The operational strain is real because AI coding tools require capacity not just for storing code, but for serving real-time assistance. Microsoft is now reportedly turning to its biggest cloud rival, Amazon, to help address capacity issues after those outages. Even without the numbers, the direction is clear: demand pressures infrastructure, infrastructure pressures uptime, and uptime pressures trust. In a tool category where engineers want the assistance in the moment, reliability can be a competitive differentiator as much as pricing.
Regulators and policy makers may sound far away from a “best month ever” internal meeting, but the structure of the market increasingly raises compliance-adjacent questions. As billing shifts to consumption and usage rises, enterprises will likely scrutinize how much AI-generated code is being produced, how it is validated, and what controls are in place. While this specific article does not mention regulators or policy changes tied to Copilot, the operational realities described here, plus the competitive urgency around pricing, can accelerate how quickly companies demand governance features and usage transparency. When costs are variable and tool use is higher, oversight tends to become a business requirement, not a “nice-to-have.”
For executives watching this, the takeaway is that GitHub is treating pricing not as paperwork, but as product strategy. One change on June 1 aligned Copilot with how rivals that charge based on consumption behave in the market. The result, per Vladimir Fedorov, was June becoming GitHub’s best month ever, alongside a surge in Copilot usage. In a race where Cursor, OpenAI’s Codex, and Anthropic’s Claude Code are all pushing developers to spend more time with AI inside the editor, GitHub’s ability to convert pricing into adoption could determine whether it keeps its footing or gets outpaced in day-to-day engineering workflows.
This story's Key Insights and Take-aways are locked.
Create a free account to unlock Executive Actions for one credit.
Register to UnlockAlways free for Executives Club members. Join the Club
More in Technology

John Carmack apologizes for Quake burnout after Sandy Petersen said it “ruined id Software”
The 30th anniversary spark turned into a rare founder-to-founder reckoning on incentives, intensity, and a “Doom++” path not taken.

Alibaba’s QwenAgentWorld trains models to predict environments, not act, and boosts 7 benchmarks
Qwen-AgentWorld flips agent training on its head by learning what environments will return next, then testing transfer across seven domains.

Gemini 3.5 Flash adds screen control, and Google folds it into one agentic tool
Google built computer use into Gemini 3.5 Flash, removing the need for a separate model and pushing enterprises to decide fast.
