AI productivity gains remain unproven despite massive infrastructure investments

While massive investments are being poured into artificial intelligence, evidence of broad productivity gains remains limited outside of software development. Major technology firms including Microsoft, Meta, Amazon, and Google are projected to spend over $600 billion on AI infrastructure in 2026, contributing to a global capital expenditure on chips and data centers expected to exceed $1 trillion.

Despite high adoption rates, such as ChatGPT reaching one billion users, actual usage depth varies. Surveys indicate that while many employees use AI, daily or frequent usage is low. For example, Gallup polling shows 52% of US employees use AI at work, but only 15% use it daily. Furthermore, a survey of executives across the US, UK, Germany, and Australia found that while most use AI, they average only 1.5 hours of use per week.

To bridge the gap between investment and results, experts suggest moving toward work intelligence programs. These programs focus on business outcomes rather than just data collection, helping leaders in finance, human resources, and operations use workforce insights to drive measurable performance changes and organizational efficiency.