AI Budgeting and Cost Management: Why Spend Is Spiking and Governance Is Now Critical

The promise that AI would cut costs is colliding with reality. A new KPMG report finds that executives are aghast at their AI bills, and the problem is not going away. For AI builders, founders, and product teams, this shift means that AI budgeting and cost management is now as important as the technology itself.

What happened

KPMG surveyed 2,145 executives around the world and found that one-third had a limited understanding of AI usage costs. AI providers have moved from flat-rate pricing to usage-based models as compute costs have risen, catching many organizations off guard.

The report cites dramatic examples of runaway spend. Uber blew through its entire 2026 AI budget in just four months and has since set usage caps on internal AI tools. Another unnamed company spent $500 million on AI in a single month because employees had no limit on licenses.

OpenAI CEO Sam Altman acknowledged the trend in a recent interview, saying companies are telling him: "My company spent my entire 2026 budget in Q1. Can you make this more efficient?"

Why AI builders should care

For teams building AI products or deploying AI internally, the era of unlimited experimentation is ending. Enterprise AI spending is becoming a financial management priority, not just a technology one. Rob Fisher, global head of advisory at KPMG, put it plainly: "The real risk isn't investing in AI but doing so without cost visibility and an understanding of the economics of AI."

If you are shipping AI features or agents, your customers will increasingly demand cost transparency and controls. Builders who ignore AI cost visibility risk losing enterprise deals to competitors who can show predictable pricing and governance.

Practical implications

The shift to usage-based pricing means that traditional software budgeting models no longer work. Organizations must build capabilities to forecast, monitor, and manage AI spending.

Key actions for AI builders and product teams:

• Instrument cost tracking early. Add token counters, API call logs, and per-feature cost dashboards before you ship.
• Set usage caps and alerts. Uber's response to its budget blowout was to cap usage on AI tools. Build similar controls into your products.
• Educate stakeholders. Many executives still budget for AI like a flat subscription. Help them understand that AI behaves like compute: metered and usage-based.
• Design for efficiency. Optimize prompt length, model selection, and caching to reduce per-call costs. Every token saved is a dollar not spent.

Caveats

The KPMG survey provides a useful signal, but the specific dollar amounts cited are anecdotal. The $500 million monthly spend and Uber's budget exhaustion are single data points, not industry averages. The report does not disclose which companies or industries were surveyed beyond the global executive pool. Builders should treat these examples as warning stories rather than benchmarks.

Additionally, the article does not detail which AI providers have shifted to usage-based pricing or by how much. The trend is clear, but the magnitude varies by vendor and contract type.