The paradox at the core of the AI revolution: The faster we innovate, the slower we grow

Silicon Valley likes to talk about AI as if abundance is already here. The numbers in McKinsey’s 2026 State of AI survey are a lot less glamorous: 88% of chief financial officers at the world’s largest companies say the technology has not done much for their bottom line. That is the problem in one line. AI is spreading fast, but growth is still dragging its feet.

The pattern looks familiar. The iPhone, broadband and the PC all arrived with huge promises and real consumer enthusiasm, yet productivity growth kept thinning out. The source calls this the Innovation Paradox: the faster the technology moves, the slower the economy seems to grow. Consumer adoption has often raced ahead of enterprise change, and enterprise change has often stopped at the surface.

Take AI. Consumer usage has reached 53% of U.S. adults in barely three years since ChatGPT launched. Enterprise AI in productive deployment is only 10%. Inside companies, the gap is just as stark: 88% of organizations use AI in at least one function, but only 6% count as high performers, meaning more than 5% EBIT impact. That share has not budged from 2025. The firms getting real results are also three times more likely to have redesigned workflows end to end. Only 21% of adopters have done that.

The deeper issue is not the model, the chip or the chatbot. It is organizational drag. The article points to a combined overhead ratio that crosses 47% of GDP in the U.S. around 2000, after which GDP growth has not sustainably exceeded 2.5%. Government spending has climbed to 36% of GDP, corporate SG&A has doubled since the 1980s, and U.S. firms spend more than $2 trillion a year on compliance, according to the Competitive Enterprise Institute. AI may be smart, but it still has to work inside slow institutions.

And that is why electrification matters as a comparison. It did not just speed things up; it forced factories and firms to reorganize around it. AI can do more than earlier digital tools because it reasons, decides and creates, especially in judgment-heavy sectors like healthcare, education, public administration and construction. But the source’s punch line is blunt: technology is moving faster than institutions, and institutions are still winning the fight.