The US-based technology behemoth Microsoft has begun reducing employee access to Anthropic’s Claude Code and shifting engineers toward GitHub Copilot CLI as the increased usage of AI pushes up enterprise computing expenses.

According to reports, the tech business had previously given thousands of workers, including designers, project managers, and developers, access to Claude Code in an effort to promote experimentation with AI-assisted coding solutions.

The platform’s widespread uptake is now claimed to have resulted in a considerable rise in operational costs.

Despite the setback, Microsoft’s wider collaboration with Anthropic is still in place.

The company continues to have a multibillion-dollar Foundry contract that provides access to Claude models via Azure services, as well as cloud infrastructure assistance.

The increase in the cost of deploying generative AI technologies at scale is something that many IT businesses are struggling with.

Due to high employee usage, Uber Chief Technology Officer Praveen Neppalli Naga also reportedly stated that the firm spent its entire 2026 budget for AI coding tools in the first four months of the year.

The increasing cost burden is a result of the token-based pricing model used by major language models, where businesses are charged based on the quantity of text that AI systems process and produce.

According to Goldman Sachs, as companies use AI agents at scale throughout their operations, agentic AI systems might lead to a significant increase in token use by 2030.

The cost of running sophisticated AI models may decrease considerably over time, but Gartner, a research company, predicts that company investment in AI may still increase because newer agentic systems need much more token usage per assignment.

As part of initiatives to hasten their transition to artificial intelligence, many major technology companies, such as Cloudflare, Oracle, and Meta Platforms, revealed layoffs and restructuring initiatives earlier this month.