Microsoft has moved to restrict internal access to Anthropic’s Claude Code tool after surging usage reportedly triggered massive enterprise AI costs, raising fresh concerns over whether artificial intelligence is truly reducing operational expenses for major technology companies.
The development comes amid growing scrutiny of the financial sustainability of large-scale AI deployment across Silicon Valley, where firms have aggressively promoted AI as a cost-cutting solution capable of boosting efficiency and reducing reliance on human labour.
According to reports circulating within the tech industry, Microsoft had expanded access to Claude Code for thousands of engineers over the past six months as part of broader AI adoption efforts. Usage reportedly increased rapidly as developers embraced the tool for coding, debugging and software reviews. However, the company is now said to be scaling back most Claude-related licenses ahead of the end of June after internal spending on token-based AI usage rose sharply.
The decision is particularly notable because Microsoft is one of Anthropic’s largest strategic backers, having invested billions of dollars into the AI company while simultaneously integrating advanced AI systems into its broader enterprise ecosystem. Brandspur Tech News Desk gathered that Microsoft is now encouraging staff to transition toward its own lower-cost internal AI solutions in an effort to reduce mounting operational expenses tied to external AI platforms.
The controversy has reignited debate around the true economics of enterprise AI adoption, especially as more companies disclose rising infrastructure and compute expenses linked to generative AI systems.
Ride-hailing giant Uber Technologies is also facing similar cost pressures after heavily integrating AI coding assistants into engineering operations. Reports indicate that AI adoption among engineers accelerated significantly within months, with internal spending quickly exceeding projected annual budgets.
Executives reportedly discovered that frequent AI-assisted coding sessions generated substantial token consumption costs, with some power users allegedly spending hundreds or even thousands of dollars monthly on AI-related workflows. Industry insiders say several firms are now introducing internal tracking systems to monitor AI consumption patterns among employees as token expenses continue to rise.
The issue extends beyond software companies. NVIDIA executive Bryan Catanzaro recently acknowledged that compute expenses for advanced AI operations can exceed personnel costs in some cases, highlighting the growing financial burden associated with scaling AI systems.
Market analysts warn that while the cost of individual AI tokens may decline over time, overall enterprise spending on AI could still rise sharply as businesses deploy increasingly sophisticated autonomous AI agents that consume larger amounts of computing resources per task.
The revelations are beginning to challenge the dominant narrative that AI adoption will automatically translate into lower corporate spending and workforce reductions. Instead, many firms are now confronting a different reality in which rapid AI usage may create entirely new layers of operational costs tied to cloud computing, infrastructure expansion and large-scale model inference.
Industry estimates suggest major technology companies are expected to spend more than $700 billion collectively on AI infrastructure and data centre expansion this year alone, even as questions intensify over whether long-term returns will justify the enormous capital outlay.
