Aging commercial buildings need resilient power, cooling, and connectivity—not flashy technology pitches—to survive the AI era.

Are today’s commercial real estate technology experts offering genuine solutions, or are they selling this century’s version of Brighto: a miracle product supposedly capable of transforming underoccupied, outdated properties into intelligent and profitable buildings?

Some of these salespeople are little more than digital hucksters, peddling half-baked technologies that either fail outright or deliver only marginal benefits. Unfortunately, when those products fall short, the resulting skepticism damages the entire market, including legitimate companies offering technologies that can increase property values and attract tenants.

In my book, Location Location Connectivity, I argued that investors should not rely exclusively on their traditional business advisers, such as lawyers and accountants, to determine whether a technology investment is sound. Evaluating 21st-century technology requires a 21st-century adviser: a technologist.

You would not ask a plumber to prepare your taxes. Why, then, would you trust an accountant or lawyer to assess complex technologies and intelligent infrastructure? Most simply do not possess the necessary technical knowledge or practical experience.

Investors should also be wary of companies promising magical solutions. The sales pitches are depressingly familiar. A new commercial real estate technology is hailed as “the savior of the nation,” “the scientific marvel of the age,” or the product that “will make old buildings new.” Yet when the people selling it are asked to explain how it works—or whether it is genuinely resilient and free of a single point of failure—their performance often collapses.

Technology companies have always hired people who proclaim themselves “the best salesmen who ever sold.” We have all watched them confidently pitch products they barely understood.

The potential market for emerging technologies is also routinely exaggerated to attract prominent investors. Enormous projections can become part of an equally enormous deception, particularly when companies are trying to raise money quickly. As in previous investment bubbles, investors often rely on the wrong people to provide an honest assessment of the opportunity before them.

Today, real estate investors and investment funds are hearing a new sales mantra: Next-generation artificial intelligence data centers are the place to put their money. These facilities will require enormous investments, the argument goes, but they will also deliver enormous returns.

Will they?

I believe the market is being overstated, just as other emerging technology markets have been in the past. Exaggeration is practically a ritual whenever companies want investors to become excited about a new technology.

Marc Ganzi, CEO of DigitalBridge, has warned that although the immense opportunity surrounding digital infrastructure is attracting considerable investor interest, investors must distinguish the strongest ideas from the rest if they expect to achieve their anticipated returns.

Ganzi has also cited McKinsey & Company’s forecast that $7 trillion will need to be invested in AI data center infrastructure by 2030, adding that some analysts expect annual capital expenditures to exceed $800 billion.

Are those figures credible, or are investors buying into the marketing plan for the next Digital Brighto 2.0?

Next-generation AI data centers possess tremendous processing power and storage capacity, but they also consume vast amounts of electricity. They require resilient and redundant power systems capable of keeping them operational. Yet power represents only one part of the equation.

Once enormous quantities of data have been synthesized and converted into useful information, how quickly can that information be delivered to people elsewhere in the country—or on the other side of the world?

The digital hucksters selling the Brighto 2.0 package usually skip over that question. AI data centers also require extraordinary broadband capacity. Their connections must be redundant and resilient, with geographically diverse routes linking them to the rest of the world.

How fast is fast enough? If a facility is processing millions of data points to produce a concise report examining numerous variables and their combined effects, how much transmission capacity will it need? One hundred gigabits per second? A terabit per second? More?

Some data center developers have not paid sufficient attention to this part of the design process. More importantly, some existing data centers now being offered at steep discounts are poor candidates for next-generation AI workloads.

A building must be capable of supporting the extraordinary demands of a next-generation AI data center, and in many markets, no suitable existing facility may be available. The better option could be to construct a new purpose-built data center or develop a smaller-scale AI computing facility within a commercial office building.

Investors should be particularly cautious about purchasing a used data center built a decade ago. Many such facilities are already technologically obsolete and cannot be economically retrofitted for next-generation AI processing. They were not designed to provide the necessary power density per square foot or the cooling capacity required by increasingly powerful computing systems.

Traditional air-cooled data centers are giving way to liquid-cooled facilities capable of dissipating far greater quantities of heat. Retrofitting an older building to accommodate those demands is neither simple nor inexpensive. In many cases, it makes more sense to build from the ground up.

The same problem extends beyond data centers. Cities around the world contain vast inventories of commercial real estate that are technologically obsolete. These buildings lack the digital infrastructure needed to support the mission-critical applications on which modern corporations depend. A supposedly intelligent building cannot function intelligently when the infrastructure serving it belongs to another era.

So where is the credible roadmap to modernization?

Before prescribing a solution, building owners and investors must understand precisely what is deficient inside a particular property. Without a detailed assessment of its power, cooling, connectivity, security, and redundancy, decision-makers risk investing in a weak or fundamentally unsuitable solution. Senior executives need a rigorous technical analysis before committing capital, and they need it sooner rather than later.

Most downtown office buildings were never designed to support mission-critical corporate infrastructure. Consider the difficulty of securing redundant power from two separate substations, delivered along geographically diverse routes. The overwhelming majority of commercial buildings do not offer that capability because architects and developers were never expected to provide it.

Throughout much of the 20th century, the standard approach was a single electrical connection delivered along a single route. That design may once have been adequate for conventional office buildings, but it creates an obvious point of failure for companies whose operations depend on uninterrupted access to power and data.

The same limitation applies to network connectivity. Most commercial buildings were designed to receive a single telecommunications line linked to a single central office. If that connection fails, the building’s occupants can lose access to the systems on which their businesses rely.

As companies move deeper into the age of artificial intelligence and robotics, they will require buildings equipped with resilient, 21st-century intelligent infrastructure. Reliable power, advanced cooling, redundant broadband connections, diverse network routes, and the elimination of single points of failure will no longer be optional amenities. They will be basic requirements.

Commercial properties that continue offering 20th-century infrastructure will struggle to attract and retain tenants. Corporations whose operations depend on AI, automation, cloud computing, and continuous connectivity will seek buildings capable of protecting their mission-critical applications.

The buildings that cannot meet those demands will remain vacant, no matter how enthusiastically they are marketed. Companies need genuine intelligent infrastructure, not another bottle of Digital Brighto 2.0.