Supply constraints redefine private real-asset returns

By 2030, inference workloads will account for more than 40% of data center demand and compound at 35% a year, according to Insurance AUM Journal's midyear infrastructure outlook—the clearest number available for why private real assets have stopped moving with the cycle and started moving with a capacity gap.

The outlook argues that much of the market now faces supply constraints driven by the convergence of AI, electrification and energy security; the defining feature of this cycle, it says, is the widening gap between demand and the system's ability to deliver compute, power and connectivity.

For insurance general accounts, the response to that gap is direct: the best opportunities are assets contracted to private counterparties and backed by private demand, in areas where regulatory and affordability exposure can be mitigated, a move away from regulated cash-flow utilities toward an offtaker-based model that puts counterparty credit at the center of underwriting. An insurer should weigh the offtaker's creditworthiness as carefully as the physics of the asset.

The AI ecosystem, the report continues, has evolved from thematic to structural, and with valuations in headline sectors already reflecting that shift, the report is searching among second- and third-order beneficiaries. For an insurer, the most visible infrastructure—the data center itself—may be the least differentiated position, because power availability, interconnection timelines, equipment supply and permitting complexity will determine which projects deliver, making execution a differentiator rather than an operational detail, and fiber, the report notes, is emerging as a mission-critical backbone.

The report prefers development-led value creation in power, where interconnection and platform benefits offer more compelling entry points than buying operating assets at current valuations, and that preference puts the scarcity premium on the development queue, a more capital-intensive posture that demands a different kind of diligence. An operating wind farm has a performance history; a development project requires an opinion on the interconnection queue, the transformer supply chain and the permitting calendar, which is where the scarcity now lives.

The same logic shows up in less flashy corners of the asset class: the outlook cites water infrastructure as a beneficiary of climate volatility, with accelerating investment in supply security and system modernization in drought-prone regions, while in waste the European Commission's Landfill Directive, which targets cutting landfill disposal to 10% by 2035, is supporting regional investment opportunities. Neither sector carries the AI label, but both run on the same fuel: a mandated or climate-forced demand that does not wait for the economic cycle.

The real estate outlook from Insurance AUM Journal, led by Tony Charles, its head of research and strategy for global real assets, makes a parallel argument: performance increasingly follows demand shifts specific to each sector rather than the broad economic cycle. The recent 20% to 25% repricing in commercial real estate offers an attractive entry point, with constrained supply and higher construction costs underpinning fundamentals, while capital flows and transaction activity rebound as investors seek relative value.

Higher interest rates have increased the cost of capital and made new development less economically viable, the outlook says, which constrains new supply; pricing has corrected meaningfully since 2022 and yields sit at multi-year highs, but dispersion is widening. Sectors tied to long-term growth themes—senior housing, industrial and supply-constrained residential—are outperforming through stronger rent growth, and Charles's team is focused on assets acquired at a discount to replacement cost in supply-constrained markets, aligned with AI infrastructure, reshoring-driven logistics, aging demographics and energy-related demand. In the U.S., that means distribution assets within data center hubs and advanced manufacturing facilities supported by rising defense spending and physical AI, alongside senior housing that tracks the demographic curve.

Taken together, the two outlooks give insurers a consistent picture: the returns in both asset classes are shifting from beta to specificity, and the market is paying for the ability to identify which power grid, which region, which building type has a genuine supply gap and a credible path to completion—an underwriting skill, not an allocation decision. An insurer that treats private real assets as a spread product will end up renting the constraint; an insurer that underwrites the constraint itself gets paid to solve it.

This publication has argued that the rotation within insurers' private allocations is a move toward asset-based finance and away from consumer credit and duration-sensitive mortgage securities, rather than a retreat from the asset class. The two outlooks extend that logic into the real economy, where the next separation in private real assets will be between those who can underwrite a transformer lead time and a permitting calendar, and those who think a diversified fund is the end of the job.

The infrastructure report anchors its strategy in valuation discipline, unit economics and contracted revenue visibility, and for an insurance general account in the middle of a private-allocation review, the discipline carries a specific implication: commit capital to projects that have already cleared the bottlenecks, not to the back of the queue. The demand story is not in question; the delivery calendar is, and that calendar has become the asset class.