News decomp.dev Shuts Its Leaderboard After AI Flooding: 'Burn Tokens to Make Number Go Up' Business Accenture Caps Company-Funded Claude at $29 a Month per Employee After Internal Bills Hit $1,000 Claude Anthropic Opens Claude Opus 5.5 Fast Mode to Everyone: Up to 2.5x Output Speed at $8/$40 per Million Tokens Voice & Audio Xiaomi Open-Sources MiDashengLM-Spatial, the First End-to-End Audio Model That Knows Where Sounds Come From AI Agents Shenzhen's Xspark AI Puts Touch Inside the Model: Spark-0 Learns From 1,000 Hours of Human Tactile Data News From 15 Nodes to 389: Chinese Team Says Shanghai AI Lab Intern-S2 Helped Prove the 30-Year-Old Ghys Conjecture Security Google DeepMind and Singapore Built a Scam Detector That Runs Entirely on Your Phone, No Cloud Required Business G7 Finance Chiefs Approve Tabletop Exercises to Rehearse AI-Powered Cyberattacks on the Banking System News decomp.dev Shuts Its Leaderboard After AI Flooding: 'Burn Tokens to Make Number Go Up' Business Accenture Caps Company-Funded Claude at $29 a Month per Employee After Internal Bills Hit $1,000 Claude Anthropic Opens Claude Opus 5.5 Fast Mode to Everyone: Up to 2.5x Output Speed at $8/$40 per Million Tokens Voice & Audio Xiaomi Open-Sources MiDashengLM-Spatial, the First End-to-End Audio Model That Knows Where Sounds Come From AI Agents Shenzhen's Xspark AI Puts Touch Inside the Model: Spark-0 Learns From 1,000 Hours of Human Tactile Data News From 15 Nodes to 389: Chinese Team Says Shanghai AI Lab Intern-S2 Helped Prove the 30-Year-Old Ghys Conjecture Security Google DeepMind and Singapore Built a Scam Detector That Runs Entirely on Your Phone, No Cloud Required Business G7 Finance Chiefs Approve Tabletop Exercises to Rehearse AI-Powered Cyberattacks on the Banking System

Accenture Caps Company-Funded Claude at $29 a Month per Employee After Internal Bills Hit $1,000

Accenture Caps Company-Funded Claude at $29 a Month per Employee After Internal Bills Hit $1,000

An internal Accenture memo reported by 404 Media shows the consulting giant now centrally funds only $29 per employee per month of Claude usage, with overruns pushed onto individual projects. Heavy users were reportedly spending $500 to $1,000 monthly, and at least one employee's access was revoked entirely.

Accenture, one of the world's largest consulting firms, has quietly put a price tag on how much AI its employees get for free. According to an internal memo obtained and reported by 404 Media, the company began centrally paying only $29 per employee per month for Claude accounts as of October 1. Beyond that allowance, staff must find a specific project or client engagement willing to pick up the tab.

The memo, titled "Important Updates to Licensing and Usage Costs," stresses that all costs are charged by actual usage. In Accenture's structure, that means mapping overruns to a work breakdown structure, the internal code for a billable project. A premium tier offers a $250 weekly token allowance, but the report indicates many employees are not on it, and some found their requests to upgrade blocked.

The cap lands hard on the power users. Internal messages reviewed by 404 Media describe monthly Claude spending between $500 and $1,000 per person before the change, and at least one employee had Claude access revoked completely. Another complained in an internal channel that the cap was "honestly, this is disappointing," noting they relied heavily on Claude Code to build and test an AI accelerator being developed for multiple Accenture clients.

The memo also warns that licenses are not guaranteed: access is periodically reviewed and can be reassigned based on business priorities, role requirements, usage patterns and headcount needs. In other words, a Claude seat is now a managed, auditable cost line rather than a standing perk.

The episode is a striking data point in the enterprise AI economics debate. Accenture's most recent fiscal year closed with a record $84.5 billion in bookings, driven in significant part by AI-related work, yet the firm's own internal math suggests frontier-model seats are expensive enough that they need hard budget discipline across a workforce of well over 700,000 people. Chargebacks to client projects also quietly shift AI cost from overhead to clients, a accounting choice that could itself become a negotiating point in future contracts.

For Anthropic, the move cuts both ways. Accenture is exactly the kind of enterprise customer that consumer-facing AI revenue charts celebrate, but a $29 monthly default allowance signals that at scale, frontier-model economics still require rationing. Other large deployments will be watching whether usage-based caps like this become the template for enterprise AI procurement.

Comments (0)

Log in to join the discussion

Log In

No comments yet