AI’s next bottleneck is emerging in infrastructure and deployment, not model development, according to an entrepreneur profile. Kevin O’Leary says he would focus on helping small businesses implement AI tools or on developing data centers—arguing that demand for AI capacity is outstripping near-term construction. The report estimates that only a small amount of data center capacity is currently under construction, while the need for more compute is “insatiable.” It also highlights that the practical challenge for institutions and employers is execution: integrating AI safely into workflows and managing data. Higher education relevance is direct for campus technology leaders. If AI infrastructure investment continues to concentrate, universities may face cost and procurement pressure for compute-heavy education and research initiatives.