The Productivity Premium: What the Electrification of Factories Tells CRE Firms About AI
By Adam Gower Ph.D.
March 2026
The productivity gains from AI will not be distributed evenly across commercial real estate. History suggests they will concentrate in firms that redesign their deal cycles around what AI makes structurally possible - not in firms that use AI to do existing work faster. The electrification of American manufacturing took three decades to produce its signature productivity surge, and the delay had nothing to do with the technology. It had everything to do with whether firms were willing to reorganize around it. The same dynamic is now playing out in CRE, with one critical difference: the infrastructure already exists, the deployment friction is minimal, and the competitive pressure to act is immediate.
Key Takeaways: AI as the Third Great Utility
- AI will not distribute its productivity gains evenly across CRE - the premium will concentrate in firms that redesign their deal cycle around AI, not those that use it to accelerate existing tasks.
- Electrification took three decades to raise U.S. manufacturing productivity because factories initially used electric motors to replicate steam-powered workflows rather than redesign production around what electric power made possible.
- The Solow paradox - computers visible everywhere except in productivity statistics - resolved when businesses stopped replicating paper-based processes and began restructuring operations around digital systems. AI is at the same inflection point now.
- In CRE, the highest-leverage redesign opportunities are underwriting, lease abstraction, market research, capital formation, and portfolio monitoring - all information-dense, repetitive at scale, and currently constrained by human bandwidth.
- The productivity gains from AI may arrive faster in CRE than in prior technology cycles because digital infrastructure already exists, deployment requires no capital construction, and competitive pressure in fragmented markets accelerates adoption.
- The strategic fork is already open: firms using AI to write memos faster are on one path; firms asking what their deal process should look like given what AI makes possible are on another. The gap between those two paths will compound.
The only newsletter you need for AI in real estate from capital formation to acquisitions, operations, and exit.
The firms that win the AI transition will not be the ones that adopted AI earliest
They will be the ones that redesigned their deal cycle around it. That distinction - between adoption and redesign - is the central lesson of every prior general-purpose technology, and it applies to commercial real estate with unusual directness.
The argument advanced here is specific: AI will not distribute its productivity gains evenly across CRE. The premium will accrue to firms that restructure how they source deals, underwrite assets, form capital, and manage operations - not to firms that use AI to produce faster first drafts of the same memos they were writing before.
To understand why, it helps to examine the only historical precedent that comes close in scale: the electrification of American industry. This article is part of a broader framework for understanding AI as a foundational economic infrastructure - what I call the Utility Thesis.
Â
The productivity paradox
In 1987, economist Robert Solow observed that computers were visible everywhere except in the productivity statistics. The technology had spread across American offices. The gains had not followed. His observation - now known as the Solow paradox - turned out to have a straightforward explanation: businesses had installed computers without reorganizing their operations. They were using the new technology to do old work the old way, just with an electric typewriter that could also run spreadsheets.
The same pattern had played out six decades earlier with electricity. Edison's Pearl Street Station opened in lower Manhattan in 1882. By the early 1900s, electric motors were common in factories. And yet, for two decades, national productivity statistics barely moved. Factory operators replaced their steam engines with electric motors - and then ran the facility exactly as before. The layout was the same. The production logic was the same. Only the power source had changed.
The productivity surge came later, and it came from a different decision. Not the decision to install electric motors, but the decision to redesign the factory around what electric motors made possible.
What the factory redesign actually involved
Steam engines required centralized mechanical power. A single engine drove rotating shafts that ran the length of the building; leather belts connected each machine to the overhead line. The geometry of the factory - which machines went where, how far apart they could be, how the building itself was oriented - was determined entirely by the physics of mechanical power distribution.
Electric motors removed that constraint entirely. Each machine could have its own independent power source. The factory no longer needed to be organized around a single fixed point. For the first time, production flow could determine layout rather than the other way around. Machines could be arranged in sequence, matched to the actual order of operations. Assembly lines became mechanically viable. Floor plans expanded horizontally. Improved lighting extended operating hours and reduced error rates.
The economic results were dramatic. Between roughly 1915 and 1940, U.S. manufacturing productivity increased at approximately 5 percent annually - a pace that has rarely been matched before or since in an industrial economy. Electrified facilities consistently outperformed non-electrified competitors, with productivity differentials of 20 percent or more documented in studies of factory-level data across steel, textiles, food processing, and machinery production.
The critical observation is that these gains did not accrue to the factories that installed electric motors earliest. They accrued to the factories that used electric motors to redesign how production worked. Early adopters who simply swapped power sources remained competitive with their old peers but gained no structural advantage. The structural advantage went to the firms that asked a different question: given what electric power makes possible, what should a factory look like?
The computer parallel - and why it matters now
The Solow paradox resolved itself, eventually. Through the 1990s, businesses began to do with computers what the factory redesigners had done with electric motors: they stopped using them to replicate existing paper-based workflows and started redesigning operations around what digital systems made structurally possible.
Enterprise resource planning systems replaced fragmented departmental records with integrated data. Digital supply chains eliminated the information delays that had required companies to carry excess inventory as a buffer against uncertainty. E-commerce restructured retail distribution from the ground up. The productivity gains that materialized through the late 1990s and 2000s were not primarily a function of faster computers - they were a function of organizational redesign enabled by what computers could do.
The lag between the technology arriving and the productivity gains appearing was roughly fifteen to twenty years. That lag was not wasted time. It was the period during which businesses learned how to reorganize work around the new capability - and during which the firms that figured it out first pulled ahead of those that did not.
Artificial intelligence is now in the early phase of the AI adoption S-curve. The technology exists and its capabilities are improving rapidly. Most organizations are using it sporadically - drafting emails, summarizing documents, generating first cuts of reports. Useful. Not transformative. The redesign phase has not yet begun at scale.
What redesign means for the CRE deal cycle
Commercial real estate is an information business that has not yet been reorganized around the information capabilities now available to it. Every stage of the deal cycle - sourcing, underwriting, capital formation, asset management, disposition - involves collecting, reconciling, and interpreting data at a pace and scale constrained by human bandwidth. AI is in the process of eliminating those constraints. The question is whether CRE firms will use that elimination to do existing work faster, or to redesign what the work looks like.
Consider underwriting. The current process requires an analyst to pull rent rolls, review lease abstracts, reconcile operating statements, model assumptions, and stress-test scenarios - a process that typically takes days for a single asset. AI systems can now structure and cross-reference that data in minutes, flag anomalies against market benchmarks, and generate scenario outputs on demand. A firm that deploys this capability to speed up its existing underwriting process saves analyst time. A firm that redesigns its underwriting process around this capability - running more deals, modeling more scenarios, screening opportunities at a volume previously impossible - gains a structural sourcing and selection advantage.
Lease abstraction illustrates a similar dynamic. Extracting key terms, dates, rent escalations, and tenant obligations from a lease document is currently a paralegal task that costs time and money and introduces human error. It is also, in terms of what is actually being done, a pattern recognition exercise - exactly the class of task at which large language models are most capable. The firm that automates abstraction reduces a cost. The firm that uses automated abstraction to monitor its entire portfolio in real time - flagging lease events, tracking tenant covenant compliance, identifying renegotiation opportunities before they become urgent - has a different kind of advantage entirely.
Capital formation is where the redesign potential may be largest. Raising capital from accredited investors depends on identifying prospects, building relationships, delivering relevant content consistently over time, and maintaining those relationships across a deal cycle that may span years. The current human cost of doing this at scale is the primary constraint on how many investors a sponsor can meaningfully maintain contact with. AI can operate investor communications at a scale, personalization level, and cadence that no human team can match - not as a replacement for relationship, but as the infrastructure that makes relationship possible at volume.
Why the productivity gains may arrive faster in CRE than they did in manufacturing
Several factors suggest the lag between AI capability and AI-driven productivity gains will be shorter in CRE than the fifteen-to-twenty year computer cycle or the thirty-year electrification cycle.
First, the infrastructure already exists. Electrification required building the grid - decades of capital investment in transmission lines, substations, and distribution networks before the technology could spread. AI reaches CRE firms through software updates and cloud platforms. The deployment friction is orders of magnitude lower.
Second, the competitive pressure in CRE capital markets is intense and the field is fragmented. A mid-market sponsor competing against larger, better-capitalized firms has a strong incentive to adopt any capability that compresses the resource gap. Historically, fragmented competitive markets accelerate technology adoption because the downside of falling behind is severe and visible.
Third, the tasks that AI addresses most directly in CRE - document analysis, data reconciliation, investor communication, market monitoring - are exactly the tasks where human bandwidth is the binding constraint. The value of removing that constraint is immediate and measurable in a way that is easier to quantify than, say, the value of installing electric lighting in a 1900s factory.
The gap between potential and realized productivity still exists. But the conditions for closing it quickly are better in CRE than they were in either prior cycle.
The strategic fork
The electrification era produced a clear competitive divergence. Factories that simply replaced steam engines with electric motors captured marginal efficiency gains - enough to feel like progress, not enough to change their market position. Factories that redesigned production around electric power achieved productivity improvements that their unreconstructed competitors could not replicate, because those improvements were embedded in the firm's structure, not in any single tool.
CRE is at that fork now. That fork is also where why organizations resist AI stops being a philosophical question and becomes an operational one. The firms on one path are using AI to write offering memorandums faster, summarize documents, and draft emails. Useful. Not transformative. The firms on the other path are asking a different question: given what AI can now do, what should our deal process look like? Which tasks should humans own, which should AI own, and what becomes possible when the answer to that question is taken seriously?
The productivity premium will not be distributed across the industry. It will concentrate in the firms that answer the second question before their competitors do. That is the argument electrification makes, and it is the argument the current AI moment demands.
The only newsletter you need for AI in real estate from capital formation to acquisitions, operations, and exit.
Frequently Asked Questions
What is the productivity premium in the context of AI?
The productivity premium refers to the outsized gains that accrue to firms that redesign their workflows around AI, rather than using AI to accelerate existing processes. These gains compound over time and are difficult for competitors to replicate because they are structural rather than tool-dependent.
Why did electrification take decades to raise productivity?
Factory operators initially replaced steam engines with electric motors without changing production layouts. Productivity surged only when manufacturers redesigned facilities around what electric motors made possible - decentralized power, flexible layouts, and continuous production flows. The technology had to be absorbed organizationally, not just installed.
How does AI affect productivity in commercial real estate?
AI is automating the information-intensive tasks that currently consume analyst, paralegal, and investor relations bandwidth - underwriting, lease abstraction, market research, and capital formation outreach. Firms that restructure their deal cycles around these capabilities gain a durable productivity advantage over those that use AI only for task-level acceleration.
What is the Solow paradox and does it apply to AI?
The Solow paradox describes the lag between technology adoption and measurable productivity gains. It applied to computers in the 1980s and 1990s, and the same dynamic is visible with AI today. The resolution, historically, has been organizational redesign rather than further technological improvement.
Which CRE functions will see the largest AI productivity gains?
The highest-leverage areas are those that are information-dense, repetitive at scale, and currently dependent on significant human time:
- Underwriting - data reconciliation, scenario modeling, and deal screening
- Lease abstraction - term extraction, portfolio monitoring, and covenant tracking
- Market research - continuous signal monitoring and competitive analysis
- Investor communications - personalized outreach and relationship maintenance at scale
- Portfolio monitoring - operational anomaly detection and asset performance analysis
How long did it take for computers to produce measurable productivity gains?
The productivity gains from computing arrived approximately fifteen to twenty years after widespread adoption. They materialized when businesses stopped using computers to replicate paper-based workflows and began redesigning operations around what digital systems made structurally possible - integrated data, digital supply chains, and e-commerce.
Will AI increase GDP?
Economic projections from McKinsey, PwC, and Goldman Sachs suggest AI could add between $13 trillion and $15.7 trillion to global GDP by 2030, primarily through productivity gains rather than new product categories. These projections assume organizational adoption at scale, not merely tool-level experimentation.
What is the difference between adopting AI and redesigning around AI?
Adopting AI means using AI tools to do existing tasks faster or cheaper. Redesigning around AI means restructuring processes to take advantage of what AI makes structurally possible - running at scales, speeds, or personalization levels that were previously impossible regardless of staff size. The productivity premium belongs to the second category.
Why might AI productivity gains arrive faster than those from electrification or computing?
Three factors compress the timeline: existing digital infrastructure eliminates the need to build distribution networks; AI spreads through software updates rather than capital construction; and competitive pressure in fragmented markets like CRE creates strong incentives for early adoption. The binding constraint is organizational will, not infrastructure availability.