Full Breakdown
AI Token Pricing Uncertainty Hits Enterprises
8/12/2026, 8:02:06 AM
Why Token Costs Fluctuate
AI providers such as OpenAI and Google charge by “tokens,” the units of text processed by large language models. Unlike traditional software licences, token usage can vary dramatically with the complexity of prompts, the length of documents, and the deployment of autonomous AI agents. A simple query may cost fractions of a penny, while a multi-step analysis can run into several dollars, creating month-to-month bill swings that CFOs find difficult to budget.
Data & Statistics
- An enterprise IT director reported a 300 % rise in AI spend over six weeks after sales teams began using the tools for email drafts.
- Companies that embed AI agents into products that reach thousands of users risk “ballooning” token costs, according to academic research.
Industry Voices on Managing the Expense
- He added that token costs can become especially hard to control with “agentic processes” — “It's particularly hard when you're looking at agentic processes,” — Ventners.
- He predicts large providers will clamp down once shareholders demand profitability.
- Rob Steele, CFO of iplicit, likened vague AI prompts to sending a family member shopping without a list, emphasizing the need for precise instructions to limit spend.
- Bill Peterson, senior director of product marketing at Sumo Logic, noted that “nobody's really figured it out” regarding how to pass token costs onto customers.
Potential Pricing Paths
Software firms are experimenting with several models: raising baseline prices, charging per result, or offering “bundles” of incident-based usage. However, any approach could be upended if major language-model providers revise their own pricing, a change that would force customers to adjust budgets frequently. The industry thus faces a tension between the flexibility of token-based billing and the predictability required for enterprise budgeting.
