Sources: Meta and Microsoft are working to cut internal Claude use; Meta employees using Claude Code dropped to ~30,000 from ~60,000 earlier in 2026

The cost of using cutting-edge AI models from external providers like Anthropic has materially increased for Microsoft and Meta.

Microsoft and Meta are reportedly reducing their internal reliance on Anthropic's Claude AI models, encouraging employees to utilize their own company-developed AI tools instead. This strategic shift has led to a significant decrease in Claude usage within these tech giants. For instance, Meta has seen its employee use of Claude Code drop from approximately 60,000 users to around 30,000. Microsoft, which had planned to spend $1 billion on Anthropic's technology, has revised this figure downward by one-third. While Microsoft's spokesperson confirmed a push towards its own GitHub Copilot, employees still retain the option to use other models. This development occurs against a backdrop where businesses are finding AI spending difficult to forecast, with AI's dynamic nature contributing to unpredictable costs.

This move by Meta and Microsoft signals a broader trend of major technology companies prioritizing their internal AI development and aiming to control costs associated with third-party AI services. As these companies invest heavily in their own AI infrastructure and models, they are incentivized to steer employees towards these proprietary solutions to maximize return on investment and maintain a competitive edge. The reduction in reliance on external AI providers could also lead to greater platform lock-in for their respective ecosystems.

The shift away from Anthropic's Claude by these influential tech players could impact the competitive landscape for AI model providers. While Anthropic has captured a significant market share, a concentrated reduction in usage by major clients like Microsoft and Meta might necessitate strategic adjustments for the AI company. It also highlights the challenges businesses face in managing and forecasting AI expenditures, especially as AI tools become more integrated into core operations.