ai
Companies Set Limits on AI Token Spending After Costs Rise
Businesses are capping employee access to AI tokens following sharp increases in usage costs. Executives report shifting from encouraging broad adoption to managing budgets more tightly.
Audio version
Tap play to generate a narrated version.
Developing·Limited corroboration so far. This page will refresh as more sources emerge.
Companies across industries are introducing limits on AI token usage as spending has increased rapidly in recent months. Pylon CEO Marty Kausas said his firm approached a point where its Anthropic plan would cost more than triple its current level with 150 employees.
He decided to set ceilings for non-technical staff and instructed the company's VP of finance to determine appropriate caps. OpenAI CEO Sam Altman stated earlier this month that AI budget discussions have changed quickly. At the start of the year spending levels drew little concern, but costs are now viewed as a significant issue.
An analysis by AlphaSense found mentions of tokens in 129 earnings calls during Q2 of 2026, compared with 57 calls in the prior quarter. 84 in April. SemiAnalysis tokenomics analyst Max Kan previously supported larger token budgets, saying a $10,000 annual spend could double productivity for a $100,000 employee.
He now expresses concern that rapid policy changes may leave engineers confused. Pega CFO and COO Ken Stillwell said his company has long restricted high-cost requests rather than setting fixed caps. He noted satisfaction that more firms are now addressing the same concerns.
Some companies are tying token access to hiring decisions. MindFort CEO Brandon Veiseh said his six-person AI startup weighs token costs against additional headcount. Everlaw CTO Max Christoff supports caps paired with a process for employees to request increases when justified.
Larridin founder Russ Franklin said firms will limit which employees receive access to certain models. AI advisor Allie K. Miller reported candidates inquiring about token tiers and lab partnerships during interviews.
