A global surge in demand for data centers—driven by rapid adoption of artificial intelligence and companies increasingly outsourcing data storage—is amplifying one of the industry’s most urgent challenges: how to keep facilities cool. An outage at CME Group, the world’s largest exchange operator, has brought renewed attention to the issue. The report originated from Reuters.
CME’s popular foreign-exchange platform and futures markets across commodities, Treasuries, and equities were halted late Thursday after a cooling failure at a facility operated by Dallas-based CyrusOne. The company, which runs more than 55 data centers across the United States, Europe and Japan, said engineering teams were dispatched to the Chicago-area site on Friday to restore cooling capacity.
Data centers are dense clusters of servers and chips running continuously—processing AI models, cloud-computing workloads and high-intensity data streams. These systems draw massive amounts of electrical power and in turn produce extreme levels of heat that traditional air-cooling units often cannot handle.
“The chips that are in those data centers need to stay within certain temperatures, otherwise they either malfunction or they turn off,” said Daniel Mewton, a partner in the infrastructure, energy and natural resources practice at Slaughter and May, speaking to Reuters.
As the power demands of AI skyrocket, the risk of overheating grows, increasing the pressure on operators to rethink cooling architectures.
To manage the heat load, more operators are turning to advanced liquid-cooling technologies, which can be up to 3,000 times more efficient than air at removing heat. But liquid cooling brings its own complications, including potential leaks, corrosion risks and higher-skill maintenance requirements. It also raises concerns about water consumption.
In response, major technology companies are experimenting with new designs aimed at reducing—or eliminating—the need for external water sources. Microsoft last year unveiled a data-center design that uses zero water for cooling by recycling liquid in a closed loop between servers and chillers.
Some operators are also integrating heat-recovery systems to repurpose waste heat for nearby buildings or industrial uses.
Despite the scale of modern data centers, outages remain rare due to stringent uptime contracts. Operators often promise availability above 99.99%.
“Specific issues directly affecting cooling systems are even rarer,” Mewton said, noting that power outages are typically the more common culprit.
The pressure to expand cooling capacity is fueling a wave of industry consolidation. According to law firm White & Case, cooling can account for as much as 40% of a data center’s total energy consumption—making it a significant and rapidly growing segment of the market.
In November, power-management firm Eaton announced a $9.5 billion deal to acquire Boyd Corporation’s thermal business from Goldman Sachs Asset Management, part of a broader push to meet soaring AI-driven demand. Competitor Vertiv is also pursuing a $1 billion acquisition of PurgeRite Intermediate to boost its liquid-cooling capabilities.
As AI adoption accelerates and data-center loads intensify, the industry’s ability to innovate in cooling technologies may determine how reliably the digital economy functions.
