In a recent research report, Dario Perkins, Global Head of Macro Research at research and consulting firm TS Lombard, pointed out that according to market consensus, the year-on-year growth rate of capital expenditures by US hyperscale cloud computing vendors is nearing its peak and will slow significantly in 2027-2028. This means that even if total investment in AI infrastructure continues to grow, the pace of new spending expansion may slow, thereby weakening the momentum that previously drove rapid profit growth across the technology industry chain.
Perkins believes that in recent years, the massive capital expenditures of cloud computing giants have consistently translated into revenue for chip manufacturers, data centers, and other AI suppliers. It is estimated that currently, over 80% of AI ecosystem revenue is still related to the internal capital expenditure cycle. As investment growth slows, the driving force of this cycle on technology company revenue and profits may weaken. To maintain current profit growth, the AI industry needs more paid demand from external companies and consumers.
The report also points out that AI-related equity investment returns may become another source of profit pressure.
In 2026, some cloud computing giants will gain substantial investment returns through holding equity in other AI companies, but if the rise in technology stock valuations slows, this contribution may also decrease. Perkins therefore predicts that a slowdown in AI investment growth and reduced investment returns could jointly drag down the earnings performance of US technology companies in 2027.
