Microsoft tells engineers to use Anthropic's Claude Code Microsoft is urging thousands of its engineers to adopt Claude Code , the coding tool developed by rival Anthropic, even as the company continues to market its own GitHub Copilot as the industry standard, according to multiple reports this week. The move reflects a growing internal concern at Microsoft about Anthropic's rapid AI advances, particularly following the January 12 launch of Claude Cowork, an AI-powered desktop agent that can autonomously manage files and execute multi-step workflows on a user's computer. Internal Response to Anthropic's Challenge In the days following Cowork's release, Microsoft product leaders told colleagues the tool appeared to rival or exceed Microsoft 365 Copilot in certain capabilities, according to a report from The Information published January 28. Multiple leaders in the 365 Copilot division voiced concerns that Cowork's ability to generate Excel spreadsheets and PowerPoint presentations from various inputs could be pulling ahead of what their own product can accomplish. "Several divisions at Microsoft are moving quickly to develop AI prototypes that resemble Cowork," The Information reported, with some of those prototypes ironically powered by Anthropic's own models. CEO Satya Nadella has personally tested Clawdbot, a viral open-source AI assistant recently renamed Moltbot after a trademark dispute with Anthropic , and encouraged staff to consider developing similar products. According to the report, Nadella shared his experience with the tool in messages on an internal Teams channel called "AI Accelerator," urging employees to think about building comparable features. Broad Deployment Across Engineering Teams The adoption of Claude Code extends beyond experimentation. According to The Verge, employees from Microsoft's Experiences + Devices division, which oversees Windows, Microsoft 365 , Outlook, Bing, Edge, and Surface, were instructed last week to install Claude Code. Even non-programmers such as designers and project managers are being encouraged to explore the tool for prototyping ideas. Microsoft engineers are now expected to use both Claude Code and GitHub Copilot side by side and provide comparative feedback. The company is also crediting sales of Anthropic models toward Azure sales targets, an unusual move since Microsoft typically incentivizes salespeople only for proprietary products or OpenAI models. A Strategic Hedge, Not a Pivot Microsoft's embrace of Anthropic's tools comes alongside a deepening financial partnership between the two companies. In November, Anthropic committed to spending $30 billion on Azure cloud services, while Microsoft invested up to $5 billion in the AI startup. Anthropic models became available by default in Microsoft 365 Copilot for most commercial tenants on January 7, 2026. Despite the internal adoption of Claude Code, analysts view this as a strategic hedge rather than a shift away from OpenAI. Microsoft remains bound to a $250 billion cloud commitment from OpenAI through 2032. Microsoft also reported 15 million paid Microsoft 365 Copilot seats in its fiscal second-quarter earnings this week, with seat additions up more than 160% year over year. "I think it's a smart move that could pay big dividends in the long run," one analyst wrote. "Just like in the large language model race, there's bound to be a change in leadership at any given point in time".
