Rats, not power, could pose overlooked risk to AI data centers

Investing.com -- The artificial intelligence boom is driving a race to build data centers and secure vast amounts of electricity, but an unexpected threat to the industry's infrastructure may come from something far less sophisticated: rats.

Rodents, maintenance failures and other mundane property risks could disrupt highly automated facilities that depend on uninterrupted power and connectivity, Kade Thomas, founder and CEO of Cerv Property Solutions, said in an interview with Investing.com.

"The biggest risk is not always the things that we're talking about," Thomas said, pointing to the physical vulnerabilities that can emerge around critical infrastructure. Rats can chew through wiring and cables, potentially causing outages or damaging equipment, he said.

"Other examples include overgrown brush, dry leaf build-up, and unmanaged trees, which can obstruct HVAC air intakes, increase ambient local temperatures and create fire hazards. In hotter climates, like that of Texas, where there are 335 active data centers and 248 more planned, these relatively simple issues can quickly spiral into catastrophic damage," he added.

The warning comes as AI companies and data-center operators pour billions of dollars into expanding computing capacity, with power availability, grid connections and cooling systems increasingly viewed as constraints on growth. But Thomas said operators also need to pay greater attention to basic property maintenance, which can be overlooked amid the focus on cutting-edge technology.

A failure in routine maintenance can have an outsized impact in a data center, where even a relatively small physical problem can interrupt operations or damage expensive equipment. Pest control, inspections and preventive maintenance therefore need to be treated as part of the reliability equation rather than as secondary property services, he said.

Thomas brings a property-services perspective to the issue. Cerv Property Solutions provides services including landscaping, pool, pest, exterior, waste and janitorial work for residential and commercial properties. Thomas has spent nearly two decades in the property-services industry and previously built and sold businesses in the sector.

The broader point is that the AI infrastructure buildout is creating a new class of physical assets whose reliability depends not only on sophisticated chips and abundant electricity, but also on the less glamorous systems that keep buildings operating.

As data centers become more valuable and more densely packed with computing equipment, seemingly minor failures could become increasingly costly. A chewed cable, a neglected leak or a missed maintenance problem may not sound like an AI risk, but in a facility where uptime is critical, the consequences can be significant.

For the rapidly expanding AI industry, keeping the machines running may ultimately require paying as much attention to the basics of property management as to the technology powering the next generation of computing.

You May Also Be Interested In

• Shein targets $26-27 billion valuation in Hong Kong IPO - report
• Royal Unibrew reports higher interim first-half profit as margins improve
• ExxonMobil awards $1.1 billion in contracts for Mozambique LNG