Most Economists Doubt AI Boom Will Deliver Massive Economic Gains

The rapid expansion of artificial intelligence has triggered enormous investment in data centres, computing infrastructure and AI companies, but economists remain divided over how much economic value the technology will ultimately create. Recent analysis has highlighted concerns that the scale of AI investment may be running ahead of measurable productivity gains. The debate has intensified as technology companies continue to commit substantial sums to infrastructure while businesses experiment with ways to integrate AI into their operations.

One of the central questions is whether AI will significantly increase productivity across the wider economy. Reserve Bank of Australia Governor Michele Bullock recently said there was not yet clear evidence that AI was improving productivity in Australia. She also noted that investment in data centres and AI infrastructure was contributing to demand and inflation, while businesses may initially experience lower productivity as workers and organisations adjust to new technologies.

At the same time, some economists remain more optimistic about AI’s longer-term economic potential. A 2026 National Bureau of Economic Research study that surveyed economists, AI-company employees, policy researchers and professional forecasters found that the median respondent expected substantial improvements in AI capabilities by 2030. The economists surveyed also projected annual US GDP growth of about 2.5%, higher than baseline medium- and long-term growth forecasts from government agencies and private-sector forecasters.

Concerns about the current investment boom are nevertheless becoming more prominent. A survey of financial economists reported by NerdWallet found that more than 75% of those contacted believed the current AI investment boom showed characteristics of a bubble. Separately, Capital Economics analysts have pointed to stretched valuations, heavy market concentration and rapid growth in investment as indicators that parts of the AI-driven market may be vulnerable if expected returns fail to materialise.

The debate does not mean that economists believe artificial intelligence has no economic value. Rather, the disagreement centres on the size, timing and distribution of its benefits. AI could improve productivity in some industries while requiring significant spending on computing power, electricity, data centres and workforce adaptation. As companies continue deploying the technology, economists and policymakers will have more evidence to assess whether today’s massive AI investment produces the broad productivity gains being anticipated or whether returns prove more limited than expected.