Only 6% of Companies Are Actually Winning With AI

Just 6% of companies qualify as "AI high performers", meaning they attribute at least 5% of earnings to AI with what McKinsey calls a significant impact, according to the firm's latest State of AI survey of 1,719 professionals and business leaders, published in August 2026. That number hasn't moved at all from a year earlier, even as AI adoption itself has kept climbing.

Everyone's using it. Few are profiting from it.

McKinsey found 37% of respondents attribute at least some EBIT impact to AI, also flat compared to 2025. Put those two numbers together and you get the real story of enterprise AI this year: nearly every company has AI running somewhere, but the gap between "we use it" and "it moved the bottom line" hasn't closed at all.

It's not that AI isn't doing anything. 80% of respondents who personally use AI in their jobs say it's improved their individual productivity, and that's a hard number to argue with. The disconnect is between individual usefulness and organizational payoff. A tool that makes thousands of employees somewhat faster doesn't automatically show up as a line on an income statement, especially if nobody redesigned the workflow around it.

The agent scaling gap

There's one place real movement shows up: agentic AI. Among companies with over $1 billion in annual revenue, 40% now report scaling AI agents in production, up from 27% in the prior survey. That's a meaningful jump, and it tracks with what large enterprises have been saying all year about moving past chatbot pilots into agents that handle multi-step work. Nearly a third of respondents said their organizations chose to build agentic coding tools in-house rather than buy a vendor product, which says something about how fast this space is moving. Nobody wants to be locked into last quarter's tool.

Cost is starting to bite too. 20% of respondents cited AI-related operating costs as a real constraint on further deployment, and separately, 39% now expect their employer to cut jobs because of AI, up from 32% in 2025. McKinsey notes actual 2025 job cuts fell well short of what people predicted a year earlier, so that expectation deserves some skepticism. People have been wrong about AI-driven layoffs before.

The verdict

"88% of companies use AI" has been roughly true for a while now, and it doesn't tell you much anymore. The number that matters is 6%, the share actually converting that usage into meaningful profit, and it's stuck. If you're deciding how much more to invest in AI next year, this is the survey worth reading, not the adoption charts that just keep climbing.