Microsoft and Meta are scaling back employees’ use of AI tools developed by Anthropic, including its Claude coding assistant, while encouraging greater adoption of alternative tools within their own ecosystems. The shift was reported by The Information on October 5, 2026.
Microsoft has reduced its projected internal spending on Anthropic technology by more than one-third. The company has also been encouraging employees, particularly developers, to rely more on GitHub Copilot and other Microsoft AI tools. However, engineers can still choose to use alternative models.
Meta is making a similar transition. The number of its employees using Claude Code has reportedly fallen from approximately 60,000 earlier in 2026 to around 30,000 recently. The company is increasingly promoting its own AI-powered coding solutions as alternatives.
MetaCode, an internal coding tool, has reportedly attracted more than 30,000 employee users. Meanwhile, Muse Code, a competing coding assistant powered by Meta’s models, is being tested with external customers. According to The Information, more than 6,000 Meta employees had recently been using Muse Code.
Why does this matter?
The shift reflects a broader challenge facing large technology companies: balancing the productivity benefits of AI with the cost of deploying it across thousands of employees.
Developing and promoting internal tools can give companies greater control over their software development workflows, help manage AI spending and reduce dependence on external providers.
However, the reported changes do not mean that Microsoft and Meta are ending their relationships with Anthropic. The developments primarily concern internal employee usage and spending decisions, rather than a complete withdrawal from Anthropic’s technology.
As competition in AI coding tools intensifies, the ability to build, deploy and manage proprietary solutions could become an increasingly important advantage for major technology companies.
