The way we use and pay for AI is changing (www.techradar.com)

Recent announcements from Anthropic and GitHub are radically changing how AI services are billed, shifting from flat fees to metered usage models. Starting June 1, GitHub Copilot will implement token-based AI Credits, while Anthropic will separate agentic usage from its Claude subscription, introducing a credit pool system. This marks a significant shift away from predictable spending, now requiring businesses to monitor their usage actively in a way more akin to utility consumption than standard software purchases. Companies like Uber and Salesforce have already reported overwhelming AI expenditures, which suggests a looming reckoning for finance departments unprepared for fluctuating costs and contract implications. As organizations grapple with escalating AI expenses, many are finding that existing contracts lack key provisions for price changes, visibility over usage, and exit clauses. This situation is exacerbated by the need for robust financial oversight and clear supplier relationships, which many companies lack. The upcoming bills in August will likely bring the reality of these challenges to the forefront, prompting a reassessment of AI governance and procurement practices within organizations. The shift highlights the necessity for a deeper understanding of AI ownership costs and adequate financial management strategies to navigate this evolving landscape.