AI Infrastructure Demand Remains Robust Despite Market Rotation

A recent shift in market sentiment has seen investors rotate away from large-cap growth and technology stocks in favor of value and small-cap equities. This movement is partly driven by geopolitical instability stemming from conflict with Iran, which has bolstered energy stocks while raising fears of a potential global recession. Central to the tech sell-off is the concern that the artificial intelligence (AI) infrastructure boom has reached its zenith. With the five largest hyperscalers projected to spend approximately $700 billion on AI infrastructure this year—a sum exceeding the GDP of most nations—skeptics question the sustainability of such capital expenditures and the eventual return on investment. However, operational data suggests continued demand. Alphabet reports that its tensor processing units (TPUs) from seven to eight years ago remain fully utilized, and neocloud provider CoreWeave indicates that five-year-old GPUs are still fully booked. Although rental rates have declined by 70% from their peaks, these assets continue to generate revenue by transitioning from training to inference tasks. Further validation comes from Taiwan Semiconductor Manufacturing (TSMC), which has significantly increased its capital expenditures to expand advanced chip production. Given TSMC's critical role in the supply chain, its aggressive expansion suggests a high level of confidence in the long-term economics of the AI cloud business. For investors, this suggests that current volatility may present an entry point for growth leaders. Companies like Alphabet and Amazon continue to see strong cloud growth, while Broadcom and AMD remain positioned to benefit from the ongoing infrastructure race.