Over 80% of Companies Report No Productivity Gains from AI Despite Billions in Investment

Widespread Corporate Adoption of AI Fails to Deliver Immediate Returns

Despite billions of dollars in corporate AI investment, the technology is failing to deliver the promised productivity boost that executives and economists anticipated. A comprehensive NBER study surveying nearly 6,000 executives from firms across the US, UK, Germany, and Australia has uncovered a stark gap between AI expectation and reality: over 80% of companies report no productivity gains from AI, and nearly 90% see no AI impact on employment or output over the past three years1

1

2

2

.

While approximately 70% of businesses are actively using AI, the limited AI usage by executives tells a revealing story. Among the third of executives who reported using AI in the workplace, actual usage amounted to only 1.5 hours per week on average, with 25% of respondents not using AI at all2

2

3

3

. This minimal engagement stands in sharp contrast to the $258.7 billion that AI firms captured in global venture capital investment in 2025, representing 61% of all venture funding1

1

.

The Productivity Paradox Returns Four Decades Later

The current situation mirrors a historical pattern that economist and Nobel laureate Robert Solow identified nearly 40 years ago. Solow's paradox, named after his famous 1987 observation that "you can see the computer age everywhere but in the productivity statistics," described how the introduction of the microcomputer initially slowed workplace productivity between the 1970s and 1980s2

2

. Productivity growth dropped from 2.9% between 1948 and 1973 to just 1.1% after 1973, as information overload from new computers created administrative burdens rather than efficiencies1

1

.

Apollo chief economist Torsten Slok recently invoked this historical parallel, noting that "AI is everywhere except in the incoming macroeconomic data." He observed that outside the Magnificent 7 tech companies, there are "no signs of AI in profit margins or earnings expectations"2

2

. The pattern suggests that AI's limited effect on workplace performance may follow a J-curve theory, where initial performance dips before eventual gains materialize.

Future Optimism Persists Despite Current Reality

Despite the disappointing short-term results, executives maintain substantial expectations for the future of AI in workplace transformation. Survey respondents forecast that AI will increase productivity by 1.4%, boost output by 0.8%, and reduce headcount by 0.7% over the next three years1

1

3

3

.

The most widely adopted AI applications currently include text generation using Generative AI models like ChatGPT, visual content creation, and data processing through Machine Learning3

3

. Companies like Microsoft continue aggressive AI integration, with Microsoft's AI leadership claiming the technology could replace all white-collar jobs within 18 months1

1

.

Conflicting Research Adds to Uncertainty

Academic research presents contradictory findings about AI productivity. While a 2023 MIT report claimed AI implementation could increase worker performance by nearly 40%, a separate MIT study from 2024 found only a modest 0.5% productivity increase expected over the next decade2

2

. Nobel laureate Daron Acemoglu noted that while 0.5% "is better than zero," it remains "disappointing relative to the promises that people in the industry and in tech journalism are making."

A separate MIT report examining 300 AI deployments found that 95% of organizations implementing AI systems saw no return on their investment, despite $30-40 billion in enterprise spending on Generative AI3

3

. Only 5% of integrated AI pilots extracted significant financial value, while the majority showed no measurable profit impact.

The Federal Reserve Bank of St. Louis offered a more optimistic view, observing a 1.9% increase in excess cumulative productivity growth since ChatGPT's late-2022 introduction2

2

. However, ManpowerGroup's 2026 Global Talent Barometer revealed that while regular AI use among workers increased 13% in 2025, confidence in the technology's utility plummeted 18%, indicating persistent distrust2

2

.

History suggests this pattern could reverse. The IT boom eventually delivered a 1.5% increase in productivity growth from 1995 to 2005, following decades of stagnation2

2

. Whether AI follows a similar trajectory remains the critical question for businesses that have committed hundreds of billions to the technology, betting on returns that have yet to materialize in workplace productivity or employment data.