Research Roundup
Commercial real estate firms produce an enormous amount of research. Location Decision's CRE Research Roundup filters that work for the findings that matter to corporate real estate leaders, site selectors and economic developers — and, more importantly, asks what those findings mean for location strategy.
This week's research points toward a common theme.
Artificial intelligence may be a digital technology, but its growth is making some very physical things more important: concentrations of skilled workers, high-quality workplaces, power infrastructure and land capable of supporting increasingly demanding development.
AI could make talent clusters more important, not less
One of the more interesting findings this week comes from CBRE Research, which challenges the assumption that artificial intelligence will necessarily result in significantly fewer office jobs and less demand for office space.
In its new report, AI's Impact on Office Demand, CBRE estimates that just 5% of the office-using workforce is highly vulnerable to AI displacement, while 18% is highly adaptive. The firm expects office-using employment to increase about 0.9% annually over the next five years, compared with 0.6% for employment overall. citeturn0search0
That's interesting for the office market. It's potentially more important for location strategy.
CBRE finds that the jobs most susceptible to automation tend to be more task-oriented and remote. Work that is harder to automate tends to require more judgment and collaboration and is more likely to occur in person. The result could be less a wholesale destruction of office work than a change in what office workers do — and where companies want those workers located. citeturn0search0
There is already evidence of that geographic concentration.
Technology companies accounted for 21% of U.S. office leasing during the first half of 2026, according to CBRE, matching their highest share on record. Nearly two-thirds of that leasing occurred in the San Francisco Bay Area and Manhattan. Among major markets, San Jose and San Francisco have the office workforces CBRE considers least vulnerable to AI automation, followed by Washington, Seattle and Boston. citeturn0search0
That's worth thinking about from an economic development perspective.
For years, technology made a compelling case for geographic dispersion. Remote work strengthened it further: If talented employees could work from anywhere, companies theoretically had greater freedom to put operations in lower-cost markets.
AI could complicate that argument.
If the highest-value employees are increasingly people who can use AI to become more productive — rather than employees performing tasks AI can replace — access to adaptable talent becomes more valuable. And those workers aren't distributed evenly.
The competitive question for a market therefore isn't simply, How many office workers do we have? It is increasingly, What can those workers do?
That distinction could favor markets with deep concentrations of engineers, researchers, financial professionals and other highly skilled workers, even when their labor and real estate costs are substantially higher.
It also gives economic developers another reason to think about workforce development as something more sophisticated than head count. A labor shed with 100,000 workers isn't necessarily competitive with another labor shed containing 100,000 workers if one workforce is substantially better positioned to work alongside emerging technologies.
The Location Decision takeaway: AI may not eliminate the importance of geography. For some high-value functions, it could make concentrations of adaptable talent more important.
The data-center boom is becoming an infrastructure story
A second Cushman & Wakefield report published this week offers a useful view of what happens when digital demand collides with physical constraints.
The firm's India Data Centre Market H1 2026 report describes a market with nearly 3.9 gigawatts of projects under construction or planned. Hyderabad, Chennai and Delhi NCR are strengthening as secondary data-center markets, while Visakhapatnam is emerging as a destination for AI-focused data-center investment. citeturn0search9
India isn't the United States, and the market dynamics shouldn't simply be transplanted from one country to another.
The larger location lesson, however, travels well.
AI demand eventually has to land somewhere.
That means electricity. Transmission. Fiber. Water or alternative cooling strategies. Land. Permitting. Construction capacity. Community acceptance. And increasingly, some credible answer to the question of when all of those things will actually be available.
The remarkable thing about the AI infrastructure boom is how quickly a seemingly digital economic-development strategy becomes a physical site-readiness exercise.
A community may have hundreds of available acres and still not have a viable data-center site. A utility may have generation capacity but not the transmission infrastructure or delivery schedule required by the project. A proposed development can clear the technical requirements and still encounter political opposition that changes its timeline.
In other words, "available" is losing some of its usefulness as a description of a site.
The Location Decision takeaway: For power-intensive projects, infrastructure certainty is becoming part of the real estate product. Economic developers need to be able to describe not merely what infrastructure exists, but how much is available, when additional capacity can arrive and what stands between today's site and an operating facility.
Land isn't the commodity it sometimes appears to be
Another Cushman & Wakefield report makes a similar point from a different direction.
Building India: Land Markets Define the Next Development Frontier, released Sept. 9, examines more than 18,000 acres of transactions across 33 Indian cities between 2021 and the first quarter of 2026. Cushman estimates those transactions represent approximately 1.4 billion square feet of potential development and about $176 billion in development value. citeturn0search1turn0search3
Again, the value for a U.S. location-decision audience isn't the Indian transaction total. It's what's determining where development moves next.
Cushman finds infrastructure-led growth corridors contributing to the rise of Tier II markets. Industrial and logistics, offices and data centers are among the uses diversifying demand beyond residential development. citeturn0search3
That's a familiar pattern.
Land can look plentiful on a map while genuinely competitive development sites remain scarce.
The difference is everything attached to the acreage: transportation access, utility capacity, entitlement, environmental conditions, neighboring uses, workforce accessibility and the amount of time required to make the property construction-ready.
This distinction becomes especially important as companies demand faster speed-to-market.
A 1,000-acre greenfield site requiring years of infrastructure work isn't necessarily competing with a 200-acre site that can begin construction quickly. In some searches, they're effectively different products.
That's an important message for communities building site inventories. Acreage is a starting point, not the product.
The Location Decision takeaway: Communities shouldn't market land simply by how much they have. Increasingly, the competitive unit is the ready acre — land with enough certainty around infrastructure, entitlement and schedule for a company to confidently put it into a project timeline.
The data in action: A manufacturer goes looking for power
And sometimes a transaction provides a useful reality check on the research.
Newmark announced this week that precision manufacturer Nominal Machining leased a 21,105-square-foot industrial facility in Rancho Cordova, California. The company serves aerospace, defense, semiconductor and medical customers. citeturn1search0
The size of the transaction isn't particularly remarkable.
The requirements are.
Newmark said the facility offered heavy power and electrical infrastructure, a fully conditioned warehouse suitable for tight-tolerance machining and metrology, highway access and room for expansion. citeturn1search0
That's a small example of a much larger change in the industrial market.
Power scarcity is frequently discussed in the context of hyperscale data centers, where electrical requirements can be enormous. But data centers aren't the only projects becoming more infrastructure-intensive.
Semiconductor production, battery manufacturing, aerospace and defense production, precision machining and other forms of advanced manufacturing can require buildings with characteristics that conventional warehouse inventory doesn't provide.
That means a market can simultaneously have industrial vacancy and a shortage of industrial space.
The distinction is usable inventory.
A vacant distribution building without the required power, floor loading, clear height, conditioning or other infrastructure doesn't necessarily solve the needs of an advanced manufacturer. Retrofitting it can add cost and, perhaps more damagingly, time.
For economic developers, this creates an opportunity to get much more granular about buildings.
Don't just know the square footage.
Know the power.
Know the gas.
Know the loading.
Know the ceiling height.
Know the expansion potential.
Know what specialized infrastructure can realistically be added — and how long it will take.
A company deciding where to put production isn't leasing an abstract number of square feet. It's buying an operating timeline.
The thread connecting all four
Taken together, this week's research points toward something larger than any individual property sector.
Location decisions are becoming increasingly dependent on scarce, difficult-to-replicate inputs.
AI doesn't make talent irrelevant; CBRE's analysis suggests it may increase the premium on adaptable talent.
The data-center boom doesn't make geography irrelevant; it makes power-rich geography exceptionally valuable.
Abundant land doesn't automatically create development capacity; infrastructure and readiness determine which acreage can actually compete.
And industrial vacancy doesn't necessarily mean an advanced manufacturer can find the building it needs.
That creates a potentially important shift in economic development strategy.
For years, communities could compete partly by demonstrating abundance: acres available, workers available, buildings available.
The next competitive advantage may be certainty.
Certainty that the workforce has the skills a company needs. Certainty that electricity can be delivered. Certainty that a site can be entitled. Certainty that a building can support the operation. And certainty that the timeline presented during site selection will survive contact with reality.
Companies can find acreage in a lot of places.
Finding a location where the pieces are already aligned is considerably harder.
CRE Research Roundup is Location Decision's recurring look at newly published commercial real estate research, with a focus on what the findings mean for site selection, economic development and corporate location strategy.