Higher education researchers and administrators are facing a new competitive constraint as universities race to build AI-ready computing capacity. A new analysis frames the “data center” as the emerging infrastructure that separates research institutions into tiers—particularly for disciplines where compute is the primary bottleneck for producing research. The piece highlights escalating power demands: traditional racks often use 5–15 kilowatts, while AI-focused configurations can draw 60–100 kilowatts or more, requiring cooling approaches like liquid cooling and new facilities investment cycles. It argues that universities are making these capital decisions now, not after a recruitment loss. For research universities, the practical takeaway is that compute capacity and facility readiness are now part of the talent pipeline—from recruiting faculty and graduate students to sustaining sponsored research and indirect cost recovery.