Key takeaways
Traditional allocation frameworks have broken down. Office and retail underperform while institutions remain under-allocated.
Niche sectors function as essential infrastructure. Capital concentrates in AI, biotech, and supply chain assets.
Supply constraints create durable pricing power. Data center and service industrial vacancy stays below 3%.
Operational complexity acts as competitive moat. Manager expertise in power and logistics determines performance.
Subsector dynamics require granular analysis. Life sciences oversupply contrasts with power-constrained data centers.
Where institutional capital is actually flowing
Institutional capital is not moving toward data centers, life sciences, and modern logistics because they are novel property types. It is moving because the risk assumptions historically embedded in “core” real estate have broken down. The unresolved tension is whether allocator classification frameworks have adjusted to how dependency, resilience, and obsolescence now actually manifest inside real assets.
Capital flows are revealing a redefinition of “essential”
Institutional capital is not abandoning real estate, but it is refusing to deploy indiscriminately. Despite long-term policy targets of roughly 10.7% to real estate, actual allocations sit closer to 9.9%, reflecting sustained underweighting rather than cyclical retreat.
That gap is not explained by risk aversion alone. Office vacancy across major US markets remains near 19% after multiple quarters of negative absorption, while retail strength has concentrated narrowly in necessity-based formats rather than the sector broadly.
In contrast, capital continues to concentrate in assets that function as embedded infrastructure for digital compute, biomedical research, and supply chain throughput. Data centers, life sciences facilities, and modern logistics are being treated less as sector bets and more as essential systems—an interpretation visible in capital flows even where asset-class labels have not yet caught up.
Why traditional real estate allocation models no longer work
Valuation resets have been substantial. Private real estate values declined approximately 20–25% from mid-2022 peaks, while public equities rebounded sharply in 2024. If pricing alone were the constraint, capital would be redeploying broadly.
Instead, deployment remains selective. Even as more than two-thirds of global real estate markets are now classified in a “buy” phase, the highest level of investor conviction since 2016, capital has not returned evenly across sectors.
Traditional allocation frameworks assumed internal sector stability: office as durable income, retail as diversified consumption exposure. Those assumptions broke when tenant demand proved to be conditional. Allocators are now distinguishing assets tied to system-level necessity from those exposed to discretionary behavior, regardless of historical sector classification.
Why data centers are outperforming traditional real estate
Global data center investment surpassed $61 billion in 2025, and the majority of large institutional investors indicate plans to increase exposure. Vacancy in primary markets averages below 3%, with new developments often substantially pre-leased prior to completion.
Demand, however, is no longer the variable allocators debate. The binding constraint is power availability. In markets such as Northern Virginia, power delivery timelines now extend multiple years, creating a scarcity premium that separates assets with secured capacity from those still awaiting grid connection.
Large-scale transactions signal conviction not simply in tenant growth, but in operational control over infrastructure access. The BlackRock-Microsoft-Nvidia consortium's $40 billion acquisition of Aligned Data Centers in October 2025, and Blackstone's nearly $16 billion take-private of AirTrunk in late 2024, demonstrate that securing power-ready capacity has become a strategic priority for institutional capital.
Nuclear power solutions at sites like Susquehanna and Three Mile Island address energy constraints that now define market access more than tenant relationships. Allocators evaluating this sector increasingly view power procurement capability as a primary determinant of value.
Life sciences returns mask underwriting risk
Life sciences real estate has delivered average annual returns of approximately 12% over the past decade, materially above broader institutional real estate benchmarks (NCREIF, sector composites). That headline figure, however, conceals wide dispersion across asset types and tenant profiles.
Oversupply in core clusters such as Boston-Cambridge, the Bay Area, and San Diego has pushed vacancy materially higher. Occupancy in leading clusters has fallen from the mid-90% range in 2022 to roughly 73–80% today, with absorption uneven and prolonged.
Tenant credit dynamics compound the risk. Early-stage biotech tenants remain sensitive to venture funding cycles and potential NIH budget adjustments, while specialized lab build-outs average over $700 per square foot in fit-out cost, limiting repositioning flexibility.
Warehouse obsolescence accelerates in modern logistics
Industrial real estate performance has fractured along vintage lines. Buildings constructed prior to 2000 experienced significant negative absorption in 2024, while newer Class A facilities absorbed hundreds of millions of square feet.
Tenant behavior explains the divergence. Large occupiers such as Amazon have dropped from as many as 300 logistics properties annually during the pandemic surge to just 61 in 2025, prioritizing fewer, more efficient buildings capable of automation (company disclosures, logistics research).
Small-bay service industrial has remained notably resilient, with vacancy near 2.1% compared to approximately 7.5% across the broader industrial market, supported by third-party logistics providers now accounting for over 30% of leasing activity. Treating industrial as a uniform defensive allocation increasingly obscures obsolescence risk.
What separates competent managers from opportunists
As institutional investors increase exposure to niche sectors, manager selection has become more exacting. Many allocators are adopting a barbell approach: pairing large diversified platforms with specialist managers possessing deep subsector expertise. This is an attempt to manage underwriting risk where operational complexity is irreducible.
The diligence process has evolved accordingly.
Operational complexity acts as a filter
Data center managers are increasingly evaluated on power procurement strategy and utility relationships.
Life sciences sponsors are assessed on tenant credit frameworks informed by venture funding cycles.
Logistics managers are scrutinized for their ability to source modern assets aligned with tenant requirements.
Capabilities in these sectors are slow to build and difficult to substitute. Power access, scientific tenant underwriting, and supply-chain tenant networks function as barriers because they are time-dependent.
Allocators increasingly view this irreversibility as a form of downside protection and increasingly favor sharpshooter sponsors over generalists in specialized sectors.
Entry points and valuation discipline in heated markets
Cap rate commentary alone explains less about pricing than it once did. Industrial cap rates have widened approximately 150 basis points from 2021–22 lows, creating apparent entry points. Yet pricing dispersion within the sector has increased rather than compressed.
Data center assets with secured power continue to trade at premiums despite broader real estate correction, while life sciences assets are seeing selective cap rate expansion tied to tenant credit and lease duration. Service industrial and essential service retail are transacting near mid-6% going-in cap rates, reflecting durability rather than sector beta.
Transaction volumes rebounded in 2024 as debt markets stabilized, supported in part by increased single-asset, single-borrower securitization activity. What clears, however, is increasingly determined by asset-level constraints rather than asset-class labels.
Across sectors, pricing factors are increasingly granular:
power access
tenant credit
lease duration
building modernity
Managers who underwrite on sector averages rather than asset-level fundamentals will find the gap between projected and realized returns widening.
Bottom line
Sectors once labeled “niche” now function as core infrastructure for the digital economy, biotech innovation, and modern supply chains. The risk for institutional portfolios is not failing to gain exposure, but misreading the drivers that actually determine outcomes within those sectors.
Allocation decisions have become inseparable from execution risk. In markets defined by power constraints, tenant credit dispersion, and accelerating obsolescence, sector exposure alone explains little. What matters is whether capital is paired with operational capabilities aligned to those realities.
Collateral Partners helps managers translate operational capabilities into investor-ready narratives, ensuring that what differentiates your approach on the ground is equally clear in the materials that reach allocation committees.
