According to McKinsey’s 2026 State of AI survey, 88% of CFOs at the world’s largest companies report that AI is not yet significantly transforming their bottom line. This phenomenon, termed the "Innovation Paradox" by KKR’s Ming Lu, suggests a troubling trend: the faster we innovate, the slower we grow.
The Double K-Shaped Pattern
Technology diffusion in the digital age is following a distinct double K-shape. The first divergence is between consumers and enterprises. While consumer AI usage has surged to 53% of U.S. adults just three years after ChatGPT’s launch, enterprise AI in productive deployment stands at a mere 10%.
The second K-shape exists within the corporate world itself. Although 88% of organizations use AI in at least one function, only 6% qualify as "high performers" achieving more than a 5% EBIT impact. This 82-point gap is not a technological deficit but an organizational transformation deficit. High performers are three times more likely to have redesigned their workflows end-to-end.
The Overhead Trap
Lu highlights a metric called the Combined Overhead Ratio—capturing government spending, corporate SG&A, and regulatory compliance costs. In the U.S., this ratio crossed 47% of GDP around the year 2000. Since then, GDP growth has struggled to exceed 2.5%, roughly half the rate seen in the 1960s when overhead remained below 35%.
"Most companies deploying AI today are putting a faster engine in a horse-drawn carriage," Lu argues.
Unlike electrification, which forced factories to abandon rigid layouts for continuous-flow production, digital technologies have struggled to penetrate the "stagnant" service sectors. Healthcare, education, and public administration—which account for 50% of advanced-economy GDP—remain resistant to productivity gains because they depend heavily on human judgment and face immense organizational complexity.
The lesson of the last 50 years is that the pace of adoption matters less than its depth. Real macroeconomic impact will only arrive when AI penetrates core operations and redesigns the very institutions that currently resist it.
