Key Takeaways

  • Brookfield Private Equity CEO Anuj Ranjan argues that public markets still fail to grasp the physical scale of power and grid equipment needed to back AI models.
  • Ranjan compares software demand to smartphone hardware: despite ten years of battery improvements, newer mobile use cases still drain phones daily, and AI compute will consume power at the same relentless pace.
  • Instead of competing for crowded, capital-heavy hyperscaler facilities, Brookfield targets niche manufacturers such as heat-tracing cable makers that protect physical piping inside facilities.
  • The firm targets corporate carve-outs of unglamorous industrial units where high-margin data center protection contracts did not exist two years ago.

The Hardware Trap in Artificial Intelligence

Tech investors spend billions chasing frontier model developers and hyperscale campuses. Ranjan points out that this capital rush ignores basic physics. The bottleneck extends well beyond chips to the raw utility capacity needed to keep facilities running without blacking out local grids.

He uses a simple hardware parallel to explain why power deficits will worsen rather than stabilize. “The compute use cases are only increasing,” Ranjan told Hugh MacArthur. “If you think about your iPhone, the battery in an iPhone ten years ago is very different than a battery today, but I can tell you we're also draining our batteries every day today.” Efficiency gains never reduce net consumption because software engineers find heavier tasks to soak up every watt saved. Every token generated creates physical heat and strains transmission lines. Ranjan states flatly: “I think the world is still underestimating the amount of infrastructure and power required to do all of this.”

The Profit in Boring Ancillary Parts

When every mega-cap fund chases the same data center development platforms, valuations decouple from operational cash flows. Brookfield takes a different path. The firm hunts through cross-border corporate carve-outs to acquire mid-tier industrial and manufacturing businesses that supply physical components to builders.

Ranjan points to heat-tracing cables as proof of how ancillary suppliers capture outsized upside without tech risk. These specialized electric wires wrap around pipes and fluid conduits to prevent freezing or maintain critical operational temperatures. Two years ago, no industrial analyst viewed pipe protection as a high-growth technology asset. Today, modern compute campuses run complex liquid cooling loops and extensive backup systems that cannot afford a single frozen joint or burst line.

“The amazing opportunity that wasn't an opportunity two years ago in this business was protecting data centers,” Ranjan explained. “No one would think there would be an AI build-out tailwind to a wire that makes a pipe heat resistant or protects pipes. No one's going to expect that, but that's become a real use case.” By buying these niche businesses out of corporate parents that neglected them, Brookfield gains exposure to big tech capital expenditures while paying industrial valuations.

Why It Matters

The smart capital in the AI cycle is moving away from direct equity checks into hyperscalers, where competition compresses IRR and creates technology obsolescence risk. Value is pooling in unglamorous industrial carve-outs that control critical bottlenecks in power and cooling. Investors who buy the physical supply chain at manufacturing multiples capture software-driven volume growth without underwriting model risk.