Ninety-three percent of enterprise respondents reported exceeding their AI budgets, according to the McKinsey Enterprise AI FinOps survey fielded in May 2026 across 120 enterprise participants and 75 qualified respondents in five major industries. The finding appears in a July 2026 McKinsey Quarterly article on managing enterprise AI demand.

Spending scales sharply with deployment. AI spend rises nearly fourfold as organizations move from isolated use cases to enterprise-wide adoption, and 62 percent of organizations have moved beyond experimentation into active deployment. A majority of surveyed firms expect AI spend to increase by at least 25 percent over the next 12 months.

Visibility lags outlay. McKinsey reports that 20 to 30 percent of AI spend is often unaccounted for because investments are fragmented across cloud providers, foundation model vendors, software platforms, experimentation environments, and business-unit purchases. Only 20 to 25 percent of companies have mature AI FinOps practices in place. Token usage can vary by up to 30 times when executing the same task, which complicates forecasting.

Optimization is producing measurable returns. About a third of organizations surveyed have already captured savings of 20 to 30 percent through active optimization in the prior three months. Prompt caching cuts repeated input-token costs by up to about 90 percent for retrieval-augmented generation and agents with large, stable prefixes. Organizations with high forecasting maturity save 10 percent more on AI spend than peers on average, and sourcing and negotiation levers are delivering unit cost reductions of 10 to 20 percent.

The article was written by Pankaj Sachdeva, a senior partner in McKinsey's Philadelphia office, and Wasim Lala, a partner in the Washington, DC office, with colleagues from QuantumBlack and McKinsey's Technology and AI group.

Source: McKinsey & Company - https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-cost-of-intelligence-how-cios-can-manage-ai-demand-at-scale