Rapid AI data center construction creates operational risks from inexperienced management

AI data centers are hitting record construction of 7,481.1 MW in early 2026, but rushed builds leave owners exposed to millions in downtime from preventable failures like rodent damage or clogged filters.

Published on: Sep 05, 2026
Rapid AI data center construction creates operational risks from inexperienced management

The rush to build AI data centers has left many newly completed facilities with site problems and inexperienced management, exposing owners to operational risks that can cost millions in downtime. A single preventable failure - a rodent chewing through cabling, a clogged intake filter, a broken irrigation line - can trigger losses far exceeding the minor repair costs that could have stopped it.

"I think the risks are actually very underrated," said Kade Thomas, CEO of Cerv Property Solutions. "We're rushing projects into production and basic site care is a lower priority."

Data center construction hit a record 7,481.1 megawatts in the first half of 2026, topping the 2024 peak of 6,350 MW, according to CBRE's midyear North America data center trends report. Hyperscalers and AI occupiers competing for power and compute are driving up asking rents on extremely limited supply, with new construction leased immediately upon delivery.

Institutional investors have poured capital into data centers because the returns are substantial and the focus stays locked on upfront construction. But once the buildings are finished, owners often pull property managers from other site types - office, multifamily, industrial - who are learning data center operations on the job. "You've got managers that are learning how to manage them," Thomas said. "They've got institutional real estate firms that are just trying to throw these things up as fast as possible. You've got AI companies that need these things, but everyone kind of forgets about [operations]."

Cooling complexity and the new asset class problem

Although data centers have existed for decades, the facilities going up today represent what Thomas calls a new asset class. Their complexity - particularly around cooling - has no deep bench of experienced operators comparable to other commercial real estate verticals. Data center owners, chip makers, and utilities now collaborate on thermal management systems designed for AI computing's escalating demands, each with its own water management requirements.

Cooling-related risks extend beyond equipment failure. A recent maintenance operation at a Meta data center in Wyoming drew attention over alleged bacterial contamination, and Legionella concerns in New York City have put cooling towers under renewed scrutiny. Law firm Baker Donelson notes that operators face mounting pressure around cooling-system monitoring, microbial control, maintenance protocols, and wastewater management. System cleaning and flushing activities carry particular risk if discharge pathways and treatment requirements are not fully understood before maintenance begins.

Where the real losses come from

Fire is the leading risk-cost driver for data centers, accounting for over 50% of losses, according to an August 12 report by Allianz. But the most frequent cause of claims is water damage. Other common claim triggers include wilful acts, fire, and equipment breakdowns. In hyperscale facilities, losses from damage to external cooling systems and hot works-related fire damage can each run between $50 million and $100 million.

Facilities are also growing larger and more interdependent. Hyperscale and colocation campuses bring together multiple tenants, servers, utilities, and on-site infrastructure in one physical space. "A single event can therefore trigger claims across property, construction, business interruption, liability, cyber, and financial lines," Allianz said. Christian Kolbe, global head of construction claims at Allianz Commercial, emphasized that resilience must be designed in from the earliest planning stage, not bolted on after construction.

Thomas points to easily preventable problems he sees repeatedly: broken irrigation lines that shift foundations, gaps in eaves or roofing that let rodents in, dust intake clogging filters, overgrown landscaping creating pest pathways. "Fixing a slight separation between your eave and your roof might only cost $700 or $1,000, but if a rodent gets in and chews through cabling, it could cost millions of dollars in downtime," he said.

Why this matters for real estate and construction professionals

For developers and contractors moving into the data center market, the margin between a profitable project and a loss-generating one increasingly lives in operations, not construction. Investors who treat site readiness and ongoing maintenance as afterthoughts - something facilities teams will handle later - are carrying risk they may not have priced into their models. The vendors and property managers inherited from other asset classes need training and systems built for data center environments. The firms that close the gap between construction speed and operational readiness will own a competitive advantage that compounds with every facility they bring online. For teams building their capability in this sector, understanding the full operational lifecycle has become as critical as securing power and land.


Get Daily AI News

Your membership also unlocks:

700+ AI Courses
700+ Certifications
Personalized AI Learning Plan
6500+ AI Tools (no Ads)
Daily AI News by job industry (no Ads)