Mainframes and the Race to Cool the World’s Data Trivia

As the weather heats up in the Northern Hemisphere, most of us are starting to think of ways to cool down. But humans aren’t the only ones that need to avoid over-heating. As demand for digital services surges, data centers are evolving rapidly – in size, complexity, and energy consumption. 

The rise of AI, especially generative AI, is driving massive growth in computing needs and infrastructure. But with more power comes more heat, and managing that heat is becoming one of the industry’s biggest engineering challenges. Traditional cooling methods are being pushed to their limits, while new technologies are emerging to meet the demands of next-generation workloads. 

This quiz explores some of the trends and innovations shaping the future of data center efficiency, sustainability, and thermal management. Take the quiz and then check out Penney Berryman’s article on the subject: The Race to Cool the World’s Data.

1. How much is global demand for data center capacity expected to grow annually through 2030?

 5–7%

 19–22%

 30–35%

 50%+

2. By 2030, what share of total data center demand could be for AI-ready workloads?

 20%

 35%

 50%

 Around 70%

3. What’s the estimated range of total annual data center power demand by 2030?

 60–80 GW

 100–140 GW

 171–219 GW

 300+ GW

4. Ten years ago, a 30-MW data center was considered large. What’s the new “normal” today?

 50 MW

 100 MW

 150 MW

 200 MW

5. What major physical problem increases as data centers grow in size and power?

 Bandwidth throttling

 Fiber cable fatigue

 Heat generation

 Disk fragmentation

6. Which of the following is a non-traditional but promising air-based cooling approach?

 KyotoCooling

 CRAH units

 Liquid chillers

 Air filtration towers

7. What’s one surprising benefit of KyotoCooling compared to traditional CRAH systems?

 Faster network speeds

 Up to 92% reduction in power usage

 Doubles rack capacity

 Works without external ventilation

8. Which cooling method submerges entire servers in a special fluid that evaporates and recondenses?

 Cold plate cooling

 Direct-to-chip

 Hydrothermal venting

 Immersion cooling

9. What cooling approach gets as physically close to the processor core as possible?

 Direct-to-die cooling

 Liquid chillers

 KyotoCooling

 Peltier plates

10. Why isn’t direct-to-die cooling more common today?

 It’s incompatible with AI hardware

 It has no performance benefits

 It requires highly precise engineering

 It generates more heat than it removes

 Loading …