EXPLAINER-Keeping cool: heat a key challenge for data centers and AI

Reuters
2025/11/28
EXPLAINER-Keeping cool: heat a key challenge for data centers and AI

By Supantha Mukherjee and Lucy Raitano

STOCKHOLM/LONDON, Nov 28 (Reuters) - The global boom in data centers as companies increasingly outsource information storage and ramp up use of energy-intensive artificial intelligence is creating a key challenge for the industry - how to keep cool.

An outage at the world's biggest exchange operator CME Group from late Thursday that halted trade on its popular currency platform and in futures spanning foreign exchange, commodities, Treasuries and stocks has put a spotlight on data centers overheating.

The problem was a cooling issue at data centers operated by Dallas-headquartered CyrusOne, which operates more than 55 centers in the U.S., Europe and Japan.

CyrusOne said on Friday its engineering teams were on site at the data center near Chicago to get the cooling system back online.

WHAT CAUSES THE HEAT?

High-powered AI and cloud servers crunching data need huge amounts of power, which gives off intense heat that traditional air cooling systems are often unable to cool properly.

Data centers contain racks of servers stacked together which are constantly turned on, consuming power. As they heat up, they require constant cooling.

"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 law firm Slaughter and May.

WHAT CAN DATA CENTER OPERATORS DO ABOUT IT?

More data centers are looking to use water or specialized coolants instead of air cooling, as liquid cooling can be 3,000 times more efficient than air at removing heat.

Liquid cooling however can create its own challenges, including potential leaks, corrosion and the need for specialized maintenance. It can also be water intensive.

Companies are looking to find ways to reduce outside coolants. Microsoft last year launched a new data center design that consumes zero water for cooling.

According to the company, its new technologies recycle water through a closed loop, circulating between the servers and chillers to dissipate heat without needing a fresh supply.

There are also systems to recover and reuse waste heat from data centers.

HOW COMMON ARE OUTAGES LINKED TO COOLING ISSUES?

Mewton said that in general data center outages were "extremely uncommon" because of contractual requirements for operators to keep them almost always online.

"You need to be up more than 99.99% of the time sometimes," he said.

While outages overall were fairly unusual, specific issues directly affecting cooling systems were "even rarer", Mewton said. "What I most often hear (about) is obviously power issues," he said.

A WAVE OF DEAL-MAKING FOR DATA CENTER COOLING

The global appetite for data centers has sparked a wave of deal-making across the industry as companies race to build capacity to meet the surge in power and cooling needs.

Law firm White and Case estimates that up to 40% of total energy consumption in data centers is spent on cooling them down, making it a big business.

In November, power management firm Eaton ETN.N said it would buy Boyd Corporation's thermal business from Goldman Sachs Asset Management for $9.5 billion, part of a push to cater to an AI demand surge.

Peer Vertiv VRT.N is also in a $1 billion deal for PurgeRite Intermediate to expand its liquid cooling services.

(Reporting by Supantha Mukherjee and Lucy Raitano; Additional reporting by Toby Sterling and Leo Marchandon; Writing Adam Jourdan; Editing by Jan Harvey)

((adam.jourdan@thomsonreuters.com; +54 1155446882; Reuters Messaging: adam.jourdan.thomsonreuters.com@reuters.net))

應版權方要求,你需要登入查看該內容

免責聲明:投資有風險,本文並非投資建議,以上內容不應被視為任何金融產品的購買或出售要約、建議或邀請,作者或其他用戶的任何相關討論、評論或帖子也不應被視為此類內容。本文僅供一般參考,不考慮您的個人投資目標、財務狀況或需求。TTM對信息的準確性和完整性不承擔任何責任或保證,投資者應自行研究並在投資前尋求專業建議。

熱議股票

  1. 1
     
     
     
     
  2. 2
     
     
     
     
  3. 3
     
     
     
     
  4. 4
     
     
     
     
  5. 5
     
     
     
     
  6. 6
     
     
     
     
  7. 7
     
     
     
     
  8. 8
     
     
     
     
  9. 9
     
     
     
     
  10. 10