A recent The Economist article highlights a surprising disconnect: despite massive optimism and investment in AI, real-world adoption of AI by businesses is stagnating. According to a survey by the U.S. Census Bureau, the share of American workers using AI at work recently dropped to about 11%, and usage has even fallen among larger firms.
This slowdown is concerning because for AI investments — infrastructure, chips, cloud services — to pay off and justify their scale, normal businesses must integrate AI into everyday operations. Analysts estimate that firms would need on the order of hundreds of billions of dollars in annual AI revenues to make the $5 trillion global investment on AI infrastructure worthwhile.
Part of the problem seems to be disappointment with what current AI tools deliver. Several surveys of companies show that many firms report lower-than-expected returns from AI initiatives. For example, a consultancy-industry survey found that a minority of companies have seen AI outcomes exceed expectations; most say results fell short.
All this means that although headlines continue to hype AI as a transformative force — and investors continue funding infrastructure — the economic pay-off is proving slower, more uneven, and harder to deliver than many hoped. The “AI revolution” may well be underway — but as of now, it seems more like a tentative first step rather than a sweeping transformation.
