Public Exposure names that touch this market's constraint stack — potential beneficiaries and constrained exposures. Exposure can be positive, constrained, regulated, second-order, or mixed; this is read-through, not a buy list.
Equinix has direct exposure to AI and cloud infrastructure demand through its global data center and interconnection platform. The Gridlocked question is not whether demand exists; it is whether Equinix can continue converting that demand into powered, connected, high-value capacity in constrained markets. Its interconnection ecosystem is a strength, but growth still depends on power, site availability, cooling, and capital discipline.
Digital Realty is a direct AI data center demand proxy, but Gridlocked treats it as a constrained operator: demand only matters when it can be converted into powered, connected, developable capacity.
American Tower is a second-order real estate infrastructure exposure to the AI buildout. It does not directly solve data center power or land constraints, but its tower and communications infrastructure portfolio sits near the connectivity layer that supports mobile data growth, cloud usage, edge infrastructure, and network densification. The exposure is indirect and should be treated as communications-infrastructure adjacency rather than a clean AI data center winner.
Crown Castle is an indirect connectivity infrastructure exposure to the AI and cloud buildout. Its towers, small cells, and fiber assets sit near the network layer that supports data growth and connectivity demand, but the company is not a direct data center or power bottleneck play. The Gridlocked read-through is strongest where AI, cloud, edge, and mobile data growth require denser communications infrastructure.
Quanta is a public-market way to express the grid construction side of the AI infrastructure bottleneck. If data center demand forces utilities and power markets to upgrade transmission, distribution, substations, and interconnection infrastructure, Quanta sits close to the physical work required to turn load growth into deliverable power.