The AI talent pool is graduating straight from universities to corporate labs. An interesting new paper published by the National Bureau of Economic Research documents how AI research has dramatically shifted from academia to industry over the past two decades. Using a dataset linking publication records to U.S. Census employment data, the authors track 42,000 AI researchers and show that top industry salaries have quintupled since 2001. The shift inevitably drives early-career talent toward compute-rich incumbents like Google, Meta, and Microsoft. Meanwhile, university salaries stagnated or declined. As researchers move into industry, their output shifts sharply: academic publishing dropped by roughly 63% while patents increased sixfold, signaling a move from open science toward proprietary, commercially focused innovation. Additionally, female representation has grown more in academia than in industry, and pay gaps remain substantial. The study paints a picture of a frontier innovation ecosystem increasingly shaped by corporate resources and commercial incentives. Classrooms are out; the cloud is in.