AI Vendors Shift to Token Pricing as Usage Heads Toward 120 Quadrillion Tokens
Updated
Updated · ZDNet · Jun 22
AI Vendors Shift to Token Pricing as Usage Heads Toward 120 Quadrillion Tokens
3 articles · Updated · ZDNet · Jun 22
Summary
Flat-fee AI access is giving way to token-based billing, with enterprises facing sharply higher costs as vendors end the subsidized, all-you-can-use phase.
Token demand has surged since late 2025 as stronger models and agentic workflows expanded context windows into the millions and added retries, loops and other hidden usage.
Supply limits are keeping prices elevated: executives at FinOps X said top-lab token rates have been largely flat since November 2025 because GPU, component and power shortages persist, with relief not expected until 2028.
SAP said falling unit costs have not lowered overall bills—some months still saw spend double—pushing companies to build token-level tracking for model mix, caching, use-case economics and revenue impact.
The shift is already reshaping software pricing through credits, hybrid subscriptions and pass-through charges, while raising concerns that expensive access will widen gaps between teams and users who can afford advanced AI and those who cannot.
As AI token costs soar, how can businesses prove their massive investment is actually paying off?
With the GPU market gridlocked, what new hardware could make powerful AI affordable again?
Is the AI boom creating high-skilled jobs or permanently locking out entry-level workers?
AI Spending in 2026: Managing Exploding Costs Amid Cheaper Tokens and Surging Consumption
Overview
As of mid-2026, the world of artificial intelligence faces a striking paradox: while the cost per AI operation (measured in tokens) keeps dropping, overall enterprise spending on AI is soaring. This is driven by a surge in AI adoption, with worker access jumping 50% in 2025 and more companies rapidly scaling their AI projects. As a result, organizations are shifting from simply exploring AI’s potential to rigorously managing its financial impact. The focus is now on understanding and controlling costs, as the rapid expansion of AI use brings new challenges for budgets and efficiency.