AI-driven demand is reworking property investment: where REITs can win now
AI is pushing data demand to levels the market wasn't built for. That surge is putting data centre REITs-and the owners who can support them-at the center of a structural shift across public real estate.
A recent white paper from Hazelview Investments outlines the trend: AI workloads are scaling far faster than traditional compute, creating a clear supply-demand gap. McKinsey expects corporate AI investment to climb sharply through 2030, accelerating this imbalance and pulling forward real estate needs tied to power, cooling, and network proximity.
Data centre REITs: tight supply, committed pipelines
North American primary markets showed strong net absorption and pre-leasing in the first half of 2025, with 74% of under-construction capacity already committed. Vacancy sits near historic lows, and revenue growth remains in the double digits for data centre REITs.
This isn't a short-term blip. AI workloads are growing multiple times faster than non-AI workloads, and the physical footprint required-power, land, interconnectivity-doesn't come online overnight. Expect continued pricing power where power is scarce and delivery is credible. For context on market trends, see CBRE's data center market update.
Beyond data centres: towers and 5G edge
AI-enhanced mobile apps are pushing carriers to reduce latency and boost throughput. Ericsson projects the mobile AI app market to reach roughly US$44B by 2030, while global mobile data traffic grows around 15% CAGR. That points to more 5G standalone deployments, denser edge nodes, and upgraded fiber backhaul.
For tower and small-cell owners, the near-term opportunity is in sites that shorten the path between the device and compute. Edge-capable locations with reliable power and fiber win first. See the latest Ericsson Mobility Report for traffic and latency trends.
Operational efficiency: AI as a margin lever
AI doesn't just drive demand; it trims cost. Analyses across 203 occupations suggest roughly 21% of the US workforce could be touched by AI-driven automation. REIT operators are already using AI for maintenance triage, energy optimization, capex forecasting, and tenant engagement.
The practical upside: lower response times, fewer truck rolls, tighter energy spend, and cleaner renewal pipelines. Margins improve without adding headcount.
What forward-thinking REITs can execute in the next 12 months
- Secure power first: Lock interconnection queues, substation access, and dual feeds where possible. Explore on-site generation and battery storage to de-risk delivery timelines.
- Pre-lease smartly: Structure options for higher power density, phased delivery, and expansion rights. Build flexibility into cooling and rack layouts to support AI-heavy tenants.
- Standardize delivery: Use repeatable, modular designs to compress schedules and improve cost certainty. Pre-buy long-lead gear (transformers, switchgear) where balance sheet allows.
- Engineer for efficiency: Target lower PUE and water-light designs. Consider heat reuse where district or industrial offtake is viable.
- Upgrade tower sites: Prioritize locations that can support 5G standalone and edge compute pods. Improve fiber backhaul and site power resilience.
- Deploy AI in ops: Start with high-ROI use cases-predictive maintenance, energy optimization, lease abstraction, collections risk scoring, and tenant support chat.
- Tighten risk controls: Stress-test projects for power delays, equipment lead times, and refinancing windows. Maintain scenario plans for rate shifts and submarket overbuild.
- Upskill teams: Train asset management, property ops, and leasing on practical AI tools to speed analysis and decisions. If you need a curated starting point, see Complete AI Training - courses by job.
Metrics that matter
- Data centres: Pre-leasing %, MW energized vs. delivered, absorption per submarket, time-to-power, PUE/WUE, and tenant workload mix (AI vs. general compute).
- Towers/edge: Tenants per site, 5G SA readiness, fiber latency to nearest cloud on-ramp, power resilience, and churn.
- Ops/efficiency: Opex per MW, energy cost per kWh, work order response and completion times, AI-assisted ticket share, and savings verified by M&V.
Bottom line
AI is creating sustained demand for power-rich, network-dense real assets-and it's giving owners new tools to run leaner. REITs that treat AI as both a demand engine and an efficiency lever will capture more yield and more durable cash flows.
The window is open for those who can deliver power, flexibility, and speed-then back it with disciplined operations.
Your membership also unlocks:
