Artificial intelligence is reinforcing data centers as one of the most compelling long‑term infrastructure plays, according to Michael Hochanadel, managing director and head of digital infrastructure at Harrison Street Asset Management.

Speaking on his return to Infralogic’s Crossroads podcast, Hochanadel said fears of a development bubble overlook the scale and durability of underlying demand, which was strong well before the emergence of generative AI.

“Even if you set aside AI and look at what the underlying market fundamentals have been over the last several years, it’s really been solid and consistent demand growth driven largely by the public cloud companies,” he said, pointing to how AI has accelerated an existing trend.

Hochanadel highlighted the sharp increase in compute intensity associated with AI workloads, arguing this has significant implications for power, location strategy and capital deployment across the sector. While hyperscale campuses dominate headlines, he said latency, connectivity and proximity to end users are becoming increasingly important as the market evolves from AI training toward inference.

He also pointed to a broadening tenant base that now includes AI model developers, neo‑clouds and enterprise users, alongside mounting constraints around power availability and zoning.

Despite these challenges, Hochanadel said data centers are rapidly emerging as a core real assets allocation for global investors.

Click here to listen to the full podcast episode.