Microsoft Cancels Claude Code Licenses, Pushes Engineers to Copilot CLI

Microsoft's Experiences + Devices division — the team behind Windows, Microsoft 365, Outlook, Teams, and Surface — has been told to stop using Claude Code by June 30, 2026, according to The Verge. Engineers are being directed to transition their workflows to GitHub Copilot CLI in the coming weeks.

The timing is not accidental. June 30 is the last day of Microsoft's fiscal year, and canceling Claude Code licenses going into a new financial year cuts operational costs at a strategic moment, TheStreet reports.

Rajesh Jha, Microsoft's Executive Vice President for Experiences + Devices, told staff: "Claude Code was an important part of that learning. At the same time, Copilot CLI has given us something especially important: a product we can help shape directly with GitHub for Microsoft's repos, workflows, security expectations, and engineering needs."

Microsoft first opened access to Claude Code in December 2025, inviting thousands of developers, project managers, and designers to experiment with Anthropic's AI coding tool, The Verge notes. The intent was to benchmark tools in real engineering workflows. But Claude Code proved too popular — engineers consistently chose it over Microsoft's own GitHub Copilot CLI because of what they described as a significant feature gap.

On the SWE‑bench benchmark, a standard measure of real‑world coding capability, Claude Code scored 80.8% using Opus 4.6 — a result that puts it ahead of most competing tools for complex, multi‑file engineering tasks, according to Morph. GitHub Copilot CLI, which went generally available in February 2026, offers specialized agents, background delegation, and multi‑model support, but has been playing catch‑up on deep agentic capabilities.

"While Claude Code has been a popular addition, it has also undermined Microsoft's new GitHub Copilot CLI coding tool," The Verge wrote, citing internal sources.

The Microsoft pullback cannot be understood in isolation. Uber provides the most dramatic parallel. The ride‑hailing giant burned through its entire 2026 AI budget in just four months after rolling out Claude Code to roughly 5,000 engineers, Forbes reported. Adoption climbed from 32% of engineers in February to 84% by March. By spring, 95% of Uber engineers were using AI tools monthly, and roughly 70% of committed code originated from those tools.

The per‑engineer costs tell the story. Monthly spend averaged $150 to $250 per engineer, with power users running between $500 and $2,000 per month. Uber's CTO Praveen Neppalli Naga reported spending $1,200 in a single two‑hour session, according to Forbes.

"Token‑based consumption pricing does not behave like the software line items chief financial officers know how to model, and the gap between what engineers consume and what finance teams expect is no longer hypothetical," Forbes noted.

On May 13, Anthropic announced that paying Claude subscribers would soon face a separate monthly credit meter for agent tools and third‑party integrations, billed at full API rates starting June 15, Forbes reported. For enterprise users already struggling with unpredictable costs, the new pricing model adds another layer of financial uncertainty.

Claude Code's enterprise pricing is already unusual: organizations pay a base seat fee ($20/seat/month) plus actual API token usage. That means costs scale directly with how much engineers use the tool — and the more useful it is, the more expensive it gets. GitHub Copilot Enterprise, by contrast, charges a flat $39 per seat per month with no usage surcharge, according to Cosmic JS.

The Microsoft decision is not just an internal corporate story — it signals something larger about how enterprise AI tooling is evolving in 2026. Three dynamics are now clear:

First, cost predictability matters more than raw capability. Companies with thousands of engineers cannot absorb open‑ended token costs, no matter how good the AI is. Flat‑rate pricing models, like GitHub Copilot's, become strategically important at scale.

Second, vendor control is non‑negotiable. Microsoft's stated goal of toolchain unification reflects more than internal politics. A company that ships Windows, Office, and Azure needs its AI coding tools integrated with its own repositories, security policies, and workflows.

Third, the AI coding market is fragmenting. The gaps that existed in December 2025 are closing fast. GitHub Copilot CLI now supports Claude models, specialized agents, and background delegation. The question for enterprise buyers is no longer which tool is best but which tool can be safely deployed across thousands of engineers without blowing the budget.