AI deflation in commercial real estate refers to the hypothesis that artificial intelligence will reduce costs across the CRE industry — lowering the price of services, reducing headcount, and ultimately compressing operating expenses for property owners and managers. The reality, as I’m seeing it from inside a company building AI for CRE, is considerably more complicated than the headline suggests.
The Macro Argument: Bloomberg Says AI Is Deflationary
Anna Wong, Bloomberg Economics’ chief U.S. economist, has built a compelling case that AI is structurally deflationary for the broader economy. Her models suggest AI-driven productivity gains could shave 0.5 to 1.0 percentage points off inflation over the next decade by reducing labor costs and increasing output efficiency across industries.
For the macro economy, she may well be right. AI is making software engineers more productive, automating customer service at scale, and compressing the cost of content creation. These are real deflationary forces. According to Wong’s analysis, the technology sector alone could see labor cost reductions of 15-20% by 2030 as AI handles tasks previously requiring human workers.
But when I hear people apply this thesis directly to commercial real estate, I think they’re skipping a few steps.
The Counter-Argument: AI Capex Is Inflationary for CRE
Here’s what the “AI is deflationary” crowd misses about real estate: the AI boom requires an enormous amount of physical infrastructure, and that infrastructure needs to go somewhere.
Torsten Slok, Apollo’s chief economist, has been tracking AI capital expenditure with the kind of intensity that should make CRE professionals pay attention. According to Slok’s research, the major cloud providers and AI companies collectively committed over $200 billion in AI-related capex in 2025 alone. A significant portion of that spending flows directly into real estate — data centers, power infrastructure, and cooling systems.
CBRE’s H1 2026 data center report shows North American data center absorption hit record levels, with AI-driven demand accounting for roughly 60% of new leasing activity. Vacancy rates in primary data center markets (Northern Virginia, Dallas, Phoenix, Chicago) have dropped below 3%. Construction pipelines are expanding, but power availability — not land — is the binding constraint.
That’s not deflationary for CRE. That’s a demand shock.
The Nuanced Middle: Where AI Actually Changes CRE Economics
So which is it? The honest answer is both, depending on what segment of CRE you’re looking at.
Where AI is deflationary for CRE:
- Property management operations. AI-powered maintenance scheduling, tenant communication automation, and energy optimization are reducing operating costs. JLL’s 2025 Global Real Estate Technology Survey found that early adopters of AI in property management reported 8-12% reductions in operating expenses.
- Loan and asset management. This is literally what we’re building at LoanBoss. Tasks that used to take a junior analyst two hours — abstracting a loan document, running a covenant compliance test, generating a lender report — now take minutes. The cost per unit of work is dropping fast.
- Brokerage and advisory. Market comps, underwriting models, tenant prospecting — AI is compressing the time and cost of these workflows. That puts downward pressure on fees.
Where AI is inflationary for CRE:
- Data centers. Massive new demand for powered shell, with limited supply and constrained power grids. Rents are rising 15-25% year-over-year in top markets.
- Technology infrastructure. Buildings need to be smarter, which means more sensors, more network capacity, more capital expenditure. Smart building retrofits aren’t cheap.
- Talent competition. The people who can actually implement AI in CRE — data engineers, ML engineers, product managers with domain expertise — are expensive and getting more expensive.
The Lease Obligation Problem
Adam Shapiro’s research at the San Francisco Fed on inflation shock momentum is relevant here, even though he’s not writing about real estate specifically. Shapiro’s core insight is that inflation shocks have persistent effects — they don’t just reverse when the initial cause fades.
Apply that to CRE: even if AI reduces the number of employees a company needs, the lease obligations those employees sit in don’t disappear on the same timeline. A 10-year office lease signed in 2023 doesn’t care that your AI reduced headcount by 20% in 2026. You’re paying that rent through 2033.
This is the structural mismatch that the “AI will empty offices” narrative ignores. CRE moves slowly. Leases are long. Even if AI genuinely reduces space demand, the impact on rents and property values unfolds over lease-term timescales, not technology-adoption timescales.
What I’ve observed from LoanBoss clients: the companies adopting AI aren’t giving back space. They’re redeploying it. The analyst who used to spend 60% of their time on data entry now spends that time on analysis and client interaction. The headcount stays roughly flat — the work changes.
My Take From Inside the Machine
Here’s what I believe after 18 months of building AI into a CRE product:
-
AI will reduce the cost of CRE services — debt management, property management, brokerage, advisory — by 20-40% over the next five years. That’s deflationary for service costs.
-
AI will increase demand for specific real estate asset types — data centers, lab space, manufacturing facilities for AI-adjacent hardware — in ways that are clearly inflationary.
-
The net effect on office and multifamily — the asset types most CRE professionals care about — will be modest over the medium term. Lease structures are too rigid and adoption is too gradual for a rapid repricing.
-
The real disruption isn’t price, it’s capability. The firms that adopt AI will be able to manage larger portfolios with the same team, respond to market changes faster, and make better decisions with better data. The firms that don’t will lose ground — not because their costs are higher, but because their competitors are simply better.
That’s the deflation that matters in CRE. Not the price of a service, but the competitive gap between firms that use AI and firms that don’t.
Frequently Asked Questions
Will AI reduce office demand significantly? Not in the near term. Lease structures are rigid, and companies are redeploying rather than reducing headcount. Over a 10-15 year horizon, space-per-employee may decline, but this will unfold gradually through lease expirations, not sudden vacancies.
How is AI capex affecting data center real estate? Dramatically. AI-driven demand accounted for roughly 60% of new data center leasing in H1 2026 according to CBRE, and vacancy rates in primary markets are below 3%. Power availability, not land supply, is the primary constraint on new development.
Should CRE investors view AI as a risk or opportunity? Both. AI creates demand for some asset types (data centers, logistics for AI hardware) while potentially reducing demand for others (traditional office, certain retail formats). The investors who understand which segments benefit and which face headwinds will outperform.
Is AI reducing costs for CRE debt management specifically? Yes. At LoanBoss, we’re seeing tasks that previously took hours compressed to minutes — loan abstraction, covenant testing, portfolio reporting. This reduces the cost per loan of active debt management, which means smaller teams can manage larger portfolios effectively.
What does Anna Wong’s deflationary thesis mean for CRE cap rates? If AI is structurally deflationary for the broad economy, it supports a lower-for-longer interest rate environment, which is generally positive for CRE valuations. But this effect competes with AI-driven changes in space demand and operating costs. The net impact on cap rates will vary significantly by asset type and market.
This is issue #4 of the AI Newsletter. Catch up on earlier issues here.