The rapid surge in corporate AI investment has raised concerns about overbuilding infrastructure, a dynamic with direct implications for universities that invest in compute, partnerships and AI research. Major tech players plan steep capex increases and private AI labs disclosed enormous infrastructure commitments, prompting comparisons to past tech bubbles. Analysts say the risk is that infrastructure capacity may outpace near‑term demand, creating stranded assets; yet long‑term societal adoption could still validate the spending. For campuses, volatile vendor pricing and concentrated dependence on a few cloud providers complicate procurement and research planning. University research offices and CIOs should reexamine AI procurement, stress‑test multi‑vendor strategies, and align capital plans with realistic adoption timelines to avoid stranded compute investments and ensure sustainable support for faculty and student research.