Your Fastest-Growing Cost Line Doesn't Have a GL Account Yet
Ramp says AI token spend across its customers is up 20.7x since June 2025. Finance needs to manage tokens like a utility bill, not a software seat.
Your fastest-growing cost line probably doesn't have its own GL account yet.
📰 THE NEWS
In July 2026 Ramp launched AI Token Spend Management: one dashboard that pulls usage and cost from providers such as Anthropic, OpenAI, Cursor and Gemini, broken down by team, project, person or API key, with weekly briefings and real-time alerts to stop overruns. Ramp's data shows AI token spend across its customers has grown 20.7x since June 2025. The tool is free to start, even for non-Ramp customers.
A cautionary tale from Ramp's own tracking: a routine software-library upgrade made one model start generating hidden "reasoning" tokens by default, pushing up cost and latency without anyone deciding to.
🔄 THE PIVOT
Seats are predictable. Tokens are metered like electricity, and one runaway agent loop can blow a month's budget overnight. In 2026, the corporate AI policy is becoming a cost-control policy, and finance should own it.
✅ ACTIONABLE TIP: set up AI spend governance this quarter
1. Create a dedicated GL account and cost-centre tags for AI tokens, separate from SaaS subscriptions.
2. Require one API key per team or project so every dollar is attributable.
3. Set monthly budgets with alerts at 50%, 80% and 100%.
4. Review cost per use case, not just total spend: which agents save hours and which just burn tokens?
5. Ask every month: could a smaller, cheaper model do this job just as well?
💬 Does your company know its monthly AI token spend today: yes, roughly, or no idea?
👉 Join Free at www.FinanceClub.Ai
#AISpend #TokenSpend #FinOps #SpendManagement #CostControl #CFO #AIinFinance #FinanceClubAi
Sources: Ramp press release (16 Jul 2026); SiliconANGLE; ZenML LLMOps case study on Ramp.