AI Sector Faces Infrastructure Strains and Investment Risks, Experts Warn

Industry leaders say that AI agents often create more work than they save, highlighting the need for transparent and auditable systems. At the Fortune Brainstorm Tech conference, executives pointed to frequent errors and the difficulty of validating AI‑generated output.

A Confluent‑sponsored study of German IT managers found that inadequate real‑time data infrastructure hampers AI scaling, with 70 % citing missing streams, data quality and fragmented ownership as major obstacles. Respondents stress that reliable data pipelines are a prerequisite for productive AI.

Research on AI adoption in German firms describes an "adoption spiral" versus an "erosion spiral," emphasizing that leadership behaviour and visible quick wins are critical to overcoming cultural resistance and licence under‑utilisation.

NYU finance professor Aswath Damodaran warns that a correction in the AI market could be more painful than the dot‑com bust because AI firms rely on massive, debt‑financed physical infrastructure. He argues AI lacks the strong economies of scale of pure software, and unchecked growth could threaten jobs and broader society.

Goldman Sachs’ latest report projects AI‑related capital expenditures by major cloud providers to total $5.3 trillion by 2030, approaching credit‑market saturation. Analysts note that companies such as Uber and Walmart are already imposing caps on AI spend as operating costs rise. The convergence of huge capex needs and rising debt raises systemic risk for the broader financial system.