McKinsey's QuantumBlack practice published new research in July 2026 documenting how fast enterprise AI spending is outrunning budgets. The analysis, based on a McKinsey Enterprise AI FinOps survey of 120 enterprise participants across five major industries, found that AI spend increases nearly fourfold as organizations move from isolated pilot projects to enterprise-wide deployment. That shift is already underway: 62 percent of surveyed organizations have moved past the experimentation phase into active AI deployment.

The budget picture is stark. Ninety-three percent of respondents report exceeding their AI budgets, and a majority expect spending to climb by at least 25 percent over the next 12 months. McKinsey attributes much of the overrun to fragmented purchasing, with AI costs scattered across cloud providers, foundation-model contracts, software platforms and individual business units, leaving companies unable to see 20 to 30 percent of their own AI spend at any given time.

Part of the shift is structural. Vendors are moving away from flat, seat-based software pricing toward consumption-based pricing tied to token usage and API calls, which exposes companies directly to the cost of high-usage workloads that used to be hidden inside a bundled subscription fee. McKinsey found that only 20 to 25 percent of companies currently have mature financial-operations practices in place to manage that shift.

Companies that do build disciplined cost controls are seeing real savings. About a third of organizations surveyed have already achieved 20 to 30 percent reductions in AI costs through active optimization steps, including prompt caching, which McKinsey found can cut repeated input-token costs by as much as 90 percent for retrieval-heavy and agent-based workloads.

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