Accenture limits Claude funding

Accenture has introduced a new internal policy limiting the centralized funding for Anthropic's Claude Enterprise to $29 per month per employee, as revealed by an internal service note obtained by investigative technology media 404 Media. The policy, set to take effect on October 1, 2026, reflects a shift in how the company manages the costs of using AI tools.

Rising costs drive policy change

The change comes amid rising costs tied to token-based billing for AI services. According to the internal note, Anthropic charges enterprise clients a fixed cost per license, plus a usage-based fee based on the number of tokens consumed. Once an employee exceeds the $29 limit, further use must be paid for directly by a specific project or team budget.

Impact on developers and projects

Previously, some developers incurred monthly bills of up to $1,000 per person using Claude. Accenture's new policy requires those employees to find alternative funding sources or risk having access to the tool revoked. The change aims to address excessive token consumption, encouraging developers to use efficiency tools like the /clear command or prompt cache.

Contradictions in AI strategy

Accenture’s move has created tension, especially after the company announced a major partnership with Anthropic in December 2025, aiming to train 30,000 professionals on Claude. In February 2026, management promoted AI use as a factor in promotions. The new policy, which limits access to a very low amount, has been seen by some as ironic or contradictory.

Employee and industry reactions

Accenture employees have expressed disappointment, with some reporting that the policy was implemented without adequate communication. One developer mentioned that the $29 cap would significantly impact project costs, particularly for those using Claude Code to build AI accelerators for clients. Other professionals criticized the lack of transparency in how the change was rolled out.

Industry trends in AI cost control

Analysts note that the era of free or heavily subsidized AI in enterprise settings is ending. Companies like Uber and Walmart have also introduced spending caps to manage AI costs. The broader trend suggests a move toward more strategic AI use, reserving high-cost models for essential tasks while relying on cheaper alternatives for routine work.