JLL: AI Real Estate Strategy for 2026

The collision of AI and the physical world has created a serious headache for anyone in tech real estate, including us at JLL. The problem is simple: our buildings were designed for people, not for the massive power and data appetites of artificial intelligence. This mismatch is already causing operational costs to spiral, system performance to degrade, and innovation to grind to a halt. So how do you turn these expensive, outdated physical assets into engines for your AI strategy?

The Obsolete Infrastructure Problem: When Legacy Meets AI

For years, the gold standard in commercial real estate was all about flexible floor plans and decent connectivity for office work. That world is over. The demands of a modern AI lab or a data center running large language models are on a completely different planet. We’re talking about power density needs that can rocket past 300 watts per square foot in specific zones, which makes the typical 50-75 watts in a standard office look pathetic. Your building’s HVAC, designed to keep people comfortable, is useless against racks of GPUs pumping out incredible heat. And the network, built for web browsing and email, doesn’t have the low-latency, high-bandwidth muscle needed for real-time AI processing across a distributed system.

The first symptom of this problem is operational chaos. Companies are throwing money at retrofits and piecemeal fixes that don’t really work, leading to sky-high energy bills and constant system bottlenecks. When you can’t even house the advanced AI hardware you need, you’re forced to lean on public cloud services, which can create its own set of problems with data sovereignty or latency for on-site applications. This is about creating a whole environment where AI can perform, from the power coming into the building all the way to the smart controls managing it. Without a real plan, these legacy buildings become expensive liabilities in the race for AI leadership.

What Went Wrong First: The Pitfalls of Incrementalism

The first instinct for most organizations was to tinker around the edges, and it usually created more problems than it solved. A common mistake was just beefing up power circuits in a server closet or rolling in a few portable AC units. It felt like a quick win, but it ignored the building’s total electrical capacity. You can’t run a supercomputer on a household extension cord for long. This approach often led to brownouts or, worse, the need for a massively expensive and disruptive infrastructure tear-up down the road.

Another dead end was the belief that the cloud would solve everything. While public cloud providers have incredible computing power, people quickly discovered that for certain AI jobs, it’s not a silver bullet. Data gravity is a real thing. Moving huge datasets back and forth to the cloud is slow and gets expensive fast. Plus, regulations often require sensitive data to stay on-premise. For anything that needs an instant response, think autonomous robotics on a factory floor, the round-trip to a distant data center just isn’t fast enough. Relying only on the cloud often led to sticker shock and poor performance, forcing everyone to go back and think about their physical buildings again.

The final nail in the coffin was a total lack of integrated planning. IT, facilities, and real estate teams were all working in their own worlds. The IT guys would buy a rack of new AI servers, and then the facilities team would find out they had no way to power or cool it. This disconnect caused long delays and wasted money, resulting in a completely fragmented AI workplace. For this to work, real estate strategy has to be baked into the AI plan from day one, not treated as an afterthought.

The Solution: A Strategic Portfolio Alignment for AI

Getting your real estate portfolio ready for AI requires a data-heavy strategy that pulls together infrastructure, tech, and operations. The whole game is moving from reactive fixes to a deliberate, forward-looking transformation of your physical spaces.

Step 1: Complete Infrastructure Audit and AI Readiness Assessment

First thing’s first: you have to run a brutally honest audit of your buildings to see what they can actually handle. This isn’t your standard facilities walkthrough. You need to get into the weeds of your electrical capacity, looking at everything from the grid connection down to the amperage available at the rack. We need to know the real numbers. Your cooling infrastructure needs the same fine-toothed-comb treatment, assessing not just raw BTU capacity but airflow, containment strategies, and your readiness for liquid cooling, which is becoming non-negotiable for high-density GPUs. The network audit has to check for a fiber backbone, verify internal cabling standards (is it at least Cat6a?), and map out spots for private 5G or edge compute nodes to crush latency. The assessment also needs to confirm floor loading capacity for heavy gear and review physical security for protecting very expensive IP.

A JLL report from late 2025 was a wake-up call, showing that only 15% of enterprise data centers have the cooling they’ll need for projected AI demands by 2028. This proves how urgent this first step is. If you don’t have this baseline data, your next investment is just a shot in the dark. This is also when you start modeling your future AI growth to project power and space needs 3-5 years out, which keeps you from under-building and having to do this all over again.

Step 2: Implementing Modular and Scalable Design Principles

After the audit, you can start implementing modular and scalable designs. This means you stop building monolithic, fixed infrastructure and start thinking in terms of flexible, adaptable systems. In a data center, that looks like deploying modular power distribution units (PDUs), scalable cooling like in-row or direct-to-chip liquid systems, and even pre-fab data center pods that can be dropped in place quickly. You build environments that can handle swings in AI demand without needing a full-blown construction project. For example, a building might start with a baseline of air cooling but be pre-plumbed to easily add liquid cooling loops when GPU density ramps up. It’s a smarter way to spend capital and stay agile.

Think about power. Instead of one giant, expensive uninterruptible power supply (UPS), using smaller, distributed UPS units makes maintenance easier and provides better redundancy. There’s no single point of failure that can take down your whole operation. It also lets you add power capacity incrementally. This kind of design lets you cram more racks into a smaller footprint, which is a huge win in expensive markets like Midtown Atlanta. You have to build for flexibility, not just for what you need today.

Step 3: Integrating Intelligent Building Management Systems (IBMS)

An AI-ready building uses an advanced Intelligent Building Management System (IBMS) for real-time monitoring and autonomous control. These systems are far more than glorified thermostats. They use their own AI to predict and react to changes in server heat output, power consumption, and equipment health. For an AI workload, the IBMS can dynamically crank up the cooling on a specific rack that’s working hard, optimize power flow to avoid tripping a breaker, and flag a piece of gear that’s about to fail. This kind of proactive management cuts downtime, makes equipment last longer, and can slash energy costs. A 2024 study from Siemens Building Technologies found these systems can cut data center energy use by up to 20%.

The IBMS is also your first line of defense for security, using AI-driven video analytics and biometrics to guard the expensive hardware and data. In a secure AI lab, the system can monitor who comes and goes, flag weird activity, and identify unauthorized people with facial recognition. This integrated intelligence makes the physical environment predictive and protective, creating a resilient AI workplace.

Step 4: Developing a Talent and Operational Strategy

Great tech and infrastructure are only half the story. You also need the people who can run it. A real solution requires a talent and operational strategy for this new AI-centric portfolio. You have to upskill your facilities teams so they understand data center ops, liquid cooling, and high-performance networking. You also need to hire new people who know how to maintain AI hardware and manage these complex IBMS platforms. The job of a building engineer is changing fast, demanding a mix of mechanical, electrical, and serious IT skills.

Your operational rulebook also needs a complete rewrite for the critical nature of AI infrastructure. You need strict change management, constant monitoring, and rapid-response plans for any problem. A power dip that’s a minor flicker in an office could destroy a week’s worth of progress on an AI training model, costing a fortune. The entire operational framework has to put uptime and data integrity first. This only happens when cross-functional teams from IT, facilities, and the business units are in the same room making real estate decisions together.

Measurable Results: The AI-Ready Advantage

When you get the real estate aligned with your AI strategy, you see real, measurable results that show up on the bottom line. The first and most immediate win is a big drop in operational expenditure (OpEx). By using smart systems and modular designs to optimize power and cooling, companies can cut their energy bills by 15-25% a year. That’s money that can be plowed back into more AI research or other projects. For a big tech company with data centers in the Alpharetta corridor, that’s millions of dollars in savings annually.

The better infrastructure also directly improves AI performance and reliability. With clean power, efficient cooling, and fast networks, AI models train quicker and deliver more accurate results. This means faster innovation and a real competitive advantage. Think of a finance company whose AI can now spot fraud in milliseconds instead of seconds because it’s running on optimized edge hardware in their own building. The impact on preventing losses is huge. A Gartner report from early 2026 stated that companies with purpose-built AI infrastructure get their models deployed three times faster than those trying to make do with general-purpose IT.

An AI-ready portfolio is also flexible, letting you adapt to new tech without tearing the building apart. When the next generation of AI hardware comes out, a modular design means you can upgrade with minimal downtime. That kind of agility is priceless in a field moving as fast as AI. It’s also a huge magnet for talent. The best AI engineers want to work with the best tools, and a state-of-the-art facility is a powerful recruiting weapon. Having the physical infrastructure to back up your claims of being an AI leader does wonders for your employer brand.

Finally, a properly aligned real estate portfolio lowers your risk. By getting away from outdated infrastructure and putting in solid monitoring and security, you dramatically reduce the odds of system failures or data breaches. This ensures business continuity, which is essential for any mission-critical AI application. This investment in your physical plant is really an investment in the future resilience and growth of your company in the AI era.

Getting your real estate portfolio AI-ready is a major project, but it’s a strategic necessity that will determine your capacity for growth. By auditing what you have, embracing modular and intelligent design, and building the right team, you can turn your buildings from a bottleneck into a true strategic advantage in the race for technological leadership.