The United States is racing to embed artificial intelligence into its economy, but new data from WiredScore suggests the country’s real estate may not be ready.
The infrastructure story of the AI era is still largely being told in chips, data centers, and the energy to power it.
But a quieter constraint is emerging at the city block level, where the physical demands of the technology will become so prominent that commercial real estate (CRE) owners need to act quickly to ensure their buildings aren’t left behind.
Inside the commercial offices where AI-driven businesses operate, there is a serious digital resilience problem that’s becoming more pressing by the day: many spaces are not equipped with the physical and digital infrastructure—systems that support the additional compute power, heating management and connectivity networks—needed to enable full AI integration.
Significantly, these constraints will become only more pronounced as occupiers expect landlords to be able to provide the spaces to support their AI transformation ambitions.
Indexing resilience constraint globally
WiredScore’s newly published Global Cities Resilience Index offers the clearest view yet of these constraints.
Analysing more than 1,650 office buildings across 40-plus global cities, the index assesses how well CRE can support the digital, operational, and cyber demands of modern enterprise.
Research shows 40% of enterprise software applications could include AI agents by 2026, up from less than 5% in 2024. And the companies integrating these tools into their workflow expect landlords to provide offices that support these capabilities.
In fact, more than 80% of companies plan to embed those agents into daily workflows in the near-term. For some property owners in the US, this isn’t a future-state transition play, but a very real and pressing requirement to address today.
WiredScore’s data shows many US assets carry persistent gaps in connectivity, redundancy, and cyber readiness at the precise moment occupier reliance on AI use is rapidly accelerating.
Measuring the three dimensions of resilience
WiredScore measures readiness across three distinct pillars:
- Physical resilience: The capacity to protect digital systems against power disruption and environmental risk.
- Digital resilience: The connectivity infrastructure needed to sustain continuous, data-intensive operations.
- Cyber resilience: The governance and implementation frameworks that protect increasingly connected buildings from bad-faith actors.
In isolation, each dimension matters. In combination, they define whether a building can credibly support an AI-oriented occupancy.
What the data makes clear is that the infrastructure assumptions underpinning AI deployments are not always being met.
The implications of poor connectivity on AI readiness
The traditional tolerance for patchy connectivity—an occasional dropped call, a slow upload, a brief network interruption—has quietly evaporated.
AI workloads fundamentally change that calculus.
They draw continuously on cloud services, real-time data feeds, and hybrid on-premise and remote systems. Baseline loads are higher; the tolerance for latency, lower. And where connectivity falters, a dip in AI-driven workflows doesn’t just slow day-to-day operations, but can cause irreparable business-critical damage.
WiredScore’s global data captures the gap in stark terms. Around 35% of the world’s most technologically advanced buildings still have in-building mobile performance that lags behind external network capability.
Again, a gap between occupier intention and asset readiness exists. According to a 2025 report, 88% of US businesses believe “5G is critical to optimizing the use of AI in the workplace”, and that 90% say “AI improves workplace security because it automatically detects network issues.”
But these businesses also report barriers to implementation remain; 41% say current building infrastructure issues prevent them upgrading internal mobile networks.
For US cities, where AI adoption pressure is concentrated and enterprise deployment timelines are compressed, these gaps represent an overall and growing structural risk.
Obsolescence is moving faster than markets expect
The flight-to-quality in commercial real estate is a well-established story. What the WiredScore data reframes is its underlying driver.
Historically, quality in office leasing was defined by location, design, and amenity. Those factors haven’t disappeared, but today occupier priorities—and the questions they ask of CRE operators—are shifting.
It’s becoming increasingly common for operators to face questions like: “Does this building have diverse fiber routes?”, “Are there redundant connectivity pathways?”; “What happens to our operations during an outage?”, or “Can the building’s systems withstand a cyberattack targeting its access controls or IoT layer?”
For assets that cannot answer those questions credibly, the commercial exposure is real. Buildings that fall short of digital resilience thresholds face accelerated obsolescence—not because they look dated, but because they cannot support how their occupiers need to work.
The inverse is equally significant. WiredScore-certified buildings in North America see tenants sign leases averaging nine months longer than comparable non-certified assets, while those equipped with AI-ready connectivity networks, can see vacancy rates up to 50% lower than comparable stock without them.
Taken together, these statistics paint a very clear picture: digital infrastructure investment has a direct, measurable return.
The physical constraints of AI’s next chapter
It’s worth stepping back from the software narrative for a moment.
The dominant discourse around AI focuses on model capability, compute investment, and the strategic positioning of technology platforms. That framing isn’t wrong, but it’s somewhat incomplete.
The economic value of AI depends on where it can actually run. And it runs in buildings: offices where hybrid teams operate across cloud and on-site systems; campuses where AI tools are embedded into workflows that cannot tolerate disruption; commercial assets that have, until recently, been evaluated on criteria entirely disconnected from digital performance.
Indexes, such as WiredScore’s, provide a useful corrective. It shifts the analytical lens from software ambition to physical resilience, and asks which cities—and the assets within them—are genuinely prepared for what enterprise AI requires.
The answer, for much of the US commercial building stock, is: not yet.
The landlords who close that gap first will define what office value looks like in the decade ahead. The ones who don’t will find that AI didn’t just reshape their occupiers’ operations—it reshaped the real estate market around them.
