AI Fails to Lift Profits at 88% of Large Firms, McKinsey Finds

A new McKinsey survey reveals that 88% of the world’s largest companies have yet to see a meaningful bottom-line impact from artificial intelligence, underscoring a widening gap between the technology’s consumer popularity and its enterprise payoff. Only 6% of organizations qualify as high performers achieving more than a 5% boost to EBIT, a figure unchanged from 2025, despite 88% of firms reporting AI use in at least one function.

The disparity mirrors a decades-old pattern. Consumer adoption of ChatGPT reached 53% of U.S. adults within three years, the fastest uptake of any consumer technology on record. Yet productive enterprise deployment languishes at roughly 10%. By comparison, smartphones hit 80% U.S. penetration in seven years, while enterprise resource planning systems needed 33 years to reach 57% of firms. Economists call this a double K-shaped diffusion: a chasm between consumers and businesses, and a second chasm within the corporate world between a tiny elite of high performers and the vast majority.

Those high performers are three times more likely to have redesigned workflows end-to-end, but only 21% of all AI adopters have done so. The stock market reflects the divergence: 36 AI-centric companies now account for 45% of the S&P 500’s market capitalization. Over the past three years, the index returned 76%; stripping out those AI stocks cuts the return to 32%.

The Productivity Paradox is not new. Robert Solow famously noted in 1987 that the computer age was visible everywhere but in the productivity statistics. The current wave differs in one critical respect: earlier digital tools processed, transmitted, or displayed information. AI reasons, decides, and creates — capabilities that directly target the judgment-intensive tasks dominating healthcare, education, government, and construction. These so-called Baumol sectors, where productivity gains have historically been elusive, make up roughly half of advanced-economy GDP.

Yet the barriers are organizational, not technological. A cross-country metric called the Combined Overhead Ratio — summing government spending, corporate selling-general-administrative expenses, and regulatory compliance costs — shows the U.S. crossed a 47%-of-GDP threshold around 2000. Since then, sustainable GDP growth has not exceeded 2.5%, roughly half the rate seen in the 1960s when overhead sat below 35%. Government spending now stands at 36% of GDP; corporate SG&A has doubled since the 1980s; compliance costs alone exceed $2 trillion annually, according to the Competitive Enterprise Institute. An analysis of S&P 500 sector-decades from 1985 to 2025 finds 73% of observations trapped in rising overhead and declining revenue growth.

Electrification and telephony in the early 20th century diffused at similar speeds and depths across consumers and enterprises, forcing factory redesigns and new management structures such as the multidivisional corporation. Digital technologies arrived in a service-dominated economy lacking the supply-chain linkages that allowed earlier innovations to cascade across sectors. As one veteran private-equity partner put it, most companies are putting a faster engine in a horse-drawn carriage.

Real-world friction confirms the pattern. At a major consumer firm, digital channels grew quickly but failed to scale across the larger offline business because independent distributors refused to share customer data. In healthcare, giving doctors digital tools is easy; integrating them into end-to-end clinical workflows runs into data-governance rules, compliance hurdles, and the asymmetric power of physicians over administrators. These are not outliers. Writer’s 2026 enterprise AI survey, conducted with Workplace Intelligence across 2,400 global leaders, found 79% of organizations reporting adoption challenges — a double-digit increase from 2025 — and 54% of C-suite executives saying AI adoption is tearing their company apart.

History suggests the timeline for economy-wide impact exceeds a decade, often stretching to several. AI technology advances faster than any predecessor, but the institutional redesign required to absorb it moves at the pace of human organizations: slowly, painfully, and against fierce resistance. The McKinsey and Writer surveys together signal that the innovation paradox persists not because technology stalls, but because institutions have not yet adapted.