According to Silicon Valley, artificial intelligence will usher in a glorious new era of abundance. But ask the CFOs of the world’s largest companies how AI is transforming their bottom line and, according to McKinsey’s 2026 State of AI Survey, the answer for 88% is: “Not so much.”
We’ve been here before. When the iPhone launched in 2007, experts predicted a mobile-first transformation of everything. When broadband became a mass market, Business week announced a “new economy”. When the PC colonized every office, Robert Solow joked that you could see the computer age everywhere except in productivity statistics. Over five decades, waves of digital innovation have driven dramatic consumer adoption, while continuing to diminish productivity growth.
I call this the innovation paradox: the faster we innovate, the slower our growth.
Two K-shapes, one problem
The diffusion of technology in the digital age follows what I call a double-K-shaped model. The first K-shaped is between consumers and businesses. Smartphone adoption has reached 80% of Americans in seven years. It took 33 years for enterprise resource planning systems to reach 57% of U.S. businesses. Consumer use of AI soared to 53% of U.S. adults just three years after ChatGPT launched. Enterprise AI in productive deployment represents only 10%.
The second K-shape is in the corporate space. McKinsey’s 2026 survey found that 88% of organizations use AI in at least one function, but only 6% of them qualify as “high performers” and achieve an EBIT impact greater than 5%, unchanged from 2025. This 82-point gap is not a technology gap. This is an organizational transformation deficit. The 6% who succeed are three times more likely to have redesigned their workflows from start to finish. Only 21% of all adopters did so. The stock market echoes this divergence: 36 companies in the S&P 500 AI now represent 45% of the index’s market capitalization. Over three years, S&P 500 stock returned 76%; remove AI actions and the figure drops to 32%.
Which electrification was successful
In the early 20th century, electrification and telecommunications reached consumers and businesses at similar rates and, importantly, with similar reach. Electric motors have forced factories to abandon rigid layouts for flexible, continuous production. Telephones have caused coordination costs to explode. Organizational overhauls followed: Sloan’s multidivisional structure at General Motors, Taylor’s scientific leadership. The diffusion has been both profound, reshaping the way businesses operate, and broad, impacting vertically linked industrial sectors.
Digital technologies are diffusing into a fundamentally different economy. Services now dominate GDP. A hospital or government agency does not have the supply chain links of a steel manufacturer. Economist William Baumol identified the constraint: “stagnant” service sectors resist productivity improvement because they depend on human interaction. These Baumol sectors – health care, education, public administration, construction – account for around 50% of GDP in advanced economies and remain stubbornly resistant to technological transformation.
The aerial trap
Why is enterprise adoption stagnating? Not for lack of technology, but because of the accumulated organizational complexity. I developed an international measure, the Combined Overhead Ratio, which captures public spending; corporate selling, general and administrative (SG&A) expenses; and regulatory compliance costs. In the United States, this ratio crossed the threshold of 47% of GDP around 2000. Above this threshold, GDP growth did not sustainably exceed 2.5%, about half the rate reached when overhead costs remained below 35% in the 1960s. Government spending rose to 36% of GDP. Business SG&A has doubled since the 1980s. The Competitive Enterprise Institute estimates that U.S. businesses spend more than $2 trillion annually on compliance alone.
My industry analysis of S&P 500 companies (1985-2025) shows that 73% of industry observations over a decade fall into the “overhead trap”: overhead costs are rising while revenue growth is declining. As Block’s Jack Dorsey argues, corporate hierarchy is an obsolete information routing protocol. Most companies deploying AI today are putting a faster engine in a horse-drawn carriage.
AI is different, but the barriers are not
AI is categorically different from previous digital waves. Previous technologies made it easier to process, transmit or display information. AI reasons, decides and creates. This is important because the Baumol sectors do not lack information; they require a lot of judgment and require diagnosis, assessment and customization, precisely the capabilities that AI can augment.
However, obstacles to diffusion remain. My more than 40 years of experience as a business executive and private equity partner confirms a recurring trend. In a consumer business, digitalization drove impressive growth in online channels, but failed to extend to the company’s much larger offline business because independent distributors refused to cede customer data to an integrated platform. In the health field, presenting digital tools to doctors is simple; Implementing end-to-end digitalization to improve clinical workflows is a difficult task, mixed with data governance, compliance and the asymmetric power of healthcare professionals over management.
These are not exceptional cases; instead, they are representative of the many businesses in which I have been personally involved. Writer’s 2026 Enterprise AI Survey, conducted with Workplace Intelligence among 2,400 global executives, confirms the large-scale trend: 79% of organizations report challenges in AI adoption – a double-digit increase from 2025 – and 54% of C-suite executives admit that AI adoption is “tearing their business apart.”
Depth and breadth, not rhythm
The lesson of 50 years of digital innovation is that the pace of adoption matters far less than its depth and breadth. Consumer adoption of ChatGPT has been the fastest of any consumer technology in history. This didn’t move any macroeconomic needles. What matters is whether AI penetrates core operations and reshapes workflows (depth), and whether this transformation extends beyond superstar companies to the vast sectors that dominate employment and GDP (breadth).
Historical evidence, from electrification to digital waves, supports a long timetable for adoption – typically ten years, often several decades. As a technology, AI is advancing faster than any predecessor. But the organizational overhaul needed to absorb it is happening at the pace of human institutions: slowly, painfully and despite fierce resistance. The innovation paradox persists, not because technology fails to advance, but because institutions fail to adapt. Society will only see a real impact when AI is widely adopted within businesses and across major economic sectors.
The opinions expressed in comments on Fortune.com are solely the opinions of the authors and do not necessarily reflect the opinions and beliefs of Fortune.
Gn bussni
