Microsoft is pushing thousands of employees across its Windows, Microsoft 365, Teams, Bing, Edge, and Surface divisions to adopt Anthropic's Claude Code, according to sources familiar with the matter. The directive, issued last week, applies to engineers and non-technical staff alike.
The timing is hard to ignore. GitHub Copilot, Microsoft's own AI coding tool, is used by 90% of Fortune 100 companies and has crossed 20 million all-time users. And yet here's Redmond telling its own people to try the competition.
Not optional, but not mandatory either
Frank Shaw, Microsoft's chief communications officer, framed the move as standard product evaluation. Internal emails seen by reporters tell a slightly different story: the directive was described as "strongly encouraged." Several thousand developers are now running both Claude Code and GitHub Copilot inside Visual Studio and VS Code, tasked with providing direct comparisons.
The pilot extends beyond engineers. Designers and project managers are prototyping ideas with Claude Code, testing whether non-developers can ship functional code with AI assistance. That last bit should concern anyone who assumed their job was safe because they don't write software.
Microsoft's official partnership with Anthropic makes this less surprising than it first appears. The companies announced a strategic alliance in November, with Anthropic committing to purchase $30 billion of Azure compute capacity. Microsoft invested up to $5 billion in Anthropic. Claude models are now available in Microsoft Foundry, Azure, and across the Copilot product family.
The math on Claude Code
Anthropic's coding agent has been on an unusual trajectory. Claude Code hit $1 billion in run-rate revenue just six months after general availability in May 2025. For context, that's faster than most enterprise SaaS products reach $100 million.
The product works differently than traditional code completion. Claude Code acts as a multi-step coding agent: it plans, executes, makes pull requests, and iterates on tasks like dependency upgrades or test generation. Where Copilot suggests the next line, Claude Code can handle an entire workflow.
Whether that difference matters in practice depends on who you ask. Some developers swear by Claude Code's handling of complex, multi-file changes. Others find GitHub Copilot's tighter IDE integration more useful for day-to-day work. Microsoft apparently wants data on which opinion holds up at scale.
What this signals
The internal pilot reveals Microsoft hedging its AI bets. OpenAI remains a core partner, but leadership clearly wants empirical data on alternatives. If the tests go well, Microsoft could sell Claude Code directly to Azure customers, positioning a competitor's tool alongside its own.
For Anthropic, the deal validates a strategy of going deep on enterprise. The company expects revenue somewhere between $20 billion and $26 billion by end of 2026, according to Reuters. Claude Code is a meaningful piece of that projection.
For developers watching from outside, the message is clear: the AI coding market isn't settled. Even the company that sells GitHub Copilot doesn't think the race is over.
