Microsoft is cutting off internal access to Anthropic's Claude Code for thousands of its developers, just as Uber admits it cannot find a clear return on investment after its engineers burned through an entire annual AI budget in four months.
The dual moves mark a sharp shift in how the tech industry handles the surging infrastructure costs of autonomous development tools. While Microsoft is forcing a hard migration back to its own GitHub Copilot tool, Uber is going through an internal crisis over whether spiraling token consumption is worth sacrificing engineering headcount.
Microsoft told thousands of its engineers on May 14 to stop using Claude Code, an AI coding assistant made by Anthropic, and switch to GitHub Copilot CLI, Microsoft's own AI coding tool, by June 30.
The affected engineers work on products like Windows, Microsoft 365, Teams, Outlook, and Surface, inside a division called Experiences and Devices. The story was first reported by The Verge's Tom Warren, who noted that Claude Code had become "perhaps a little too popular" inside Microsoft, with engineers choosing Anthropic's tool over Microsoft's own.
In an internal memo, Microsoft EVP Rajesh Jha explained the decision as a push toward standardizing on one tool. "Claude Code was an important part of that learning," Jha wrote. "At the same time, Copilot CLI has given us something especially important: a product we can help shape directly with GitHub."
June 30 also marks the last day of Microsoft's fiscal year. The cancellation does not end Microsoft's wider relationship with Anthropic. Claude models remain accessible through Copilot CLI and other Microsoft products, and the companies' broader investment partnership is unaffected, according to The Verge.
Damilare Odedina
Why Uber Burned Through Its 2026 AI Coding Budget in Four Months
On the other side is Uber which gave roughly 5,000 engineers access to Claude Code in December 2025. By April, CTO Praveen Neppalli Naga told The Information that the company had already exhausted its entire annual budget for AI coding tools, including Claude Code and Cursor, a rival AI coding assistant.
"I'm back to the drawing board, because the budget I thought I would need is blown away already," Naga said.
The company’s costs rose because tools like Claude Code charge based on token consumption. A token is a small unit of text the AI processes. The more complex or frequent the task, the more tokens are used, and the higher the bill.
Uber's internal leaderboards had ranked engineers by usage volume, triggering an internal culture of "tokenmaxxing" that accelerated adoption fast. Per-engineer monthly costs reportedly ran between $500 and $2,000. By April, 95% of Uber's engineers were using AI tools monthly and 70% of committed code was AI-generated, according to multiple reports.
Uber COO Andrew Macdonal SaysAI Spending Is Getting "Harder to Justify"
In a Rapid Response podcast interview published Saturday, May 23, Uber president and COO Andrew Macdonald said the company has not been able to connect its AI spending to measurable improvements in consumer products.
"That link is not there yet," Macdonald said. "It's very hard to draw a line between one of those stats and, okay, now we're actually producing 25 percent more useful consumer features."
Macdonald said the company now has to weigh AI tool costs directly against headcount decisions. "We're going to have to start talking about token consumption and the associated cost versus headcount," he said. "If you're not actually able to draw a direct line to how much useful features and functionality you're shipping to your users, that trade becomes harder to justify."
In 2025, Uber spent $3.4 billion on research and development, up 9% from the year before, according to Fortune. CEO Dara Khosrowshahi has said the company is making up for rising AI spending by hiring fewer human employees. It’s a wake-up call for the rest of the industry. The AI tools meant to augment engineers are suddenly starting to cost just as much as they do.
David Adubiina
