A new higher-education analysis argues that research universities’ ability to recruit AI-era scientists increasingly depends on data center capacity and computing infrastructure—shifting the “facility problem” into a strategy and revenue question. The piece describes how AI workloads can raise power draw dramatically, from 5–15 kilowatts per rack to 60–100+ kilowatts for serious AI configurations. It also stresses the operational constraints: conventional cooling systems often cannot handle the heat load, pushing campuses toward liquid-cooling solutions such as direct liquid cooling, rear-door heat exchangers, or full immersion systems. These investments are portrayed as time-sensitive because institutions are making capital decisions before they feel recruitment losses. The analysis ties compute capacity to downstream outcomes—graduate student attraction, sponsored research activity, indirect cost recovery, and ultimately institutional mission execution—framing data center investment as part of a competitive academic position rather than a back-office upgrade.