TECH COLUMN

0-8. Why Heat Problems Are Becoming a Matter of National Competition ― In the AI Era, the Truly Strong Countries May Be Those That Can Process Heat ―

Heat

release:
0-8. Why Heat Problems Are Becoming a Matter of National Competition ― In the AI Era, the Truly Strong Countries May Be Those That Can Process Heat ―

Today, the AI race is intensifying around the world.
The United States. China. Europe. The Middle East. India.

Countries have begun investing enormous sums in
•AI
•Semiconductors
•Data centers
•EVs
•Robots
At first glance, this looks like a competition of intelligence.
But is that really the case?

■ AI Looks Like an "Intelligence Industry," but It Is Also a Resource Industry

Generative AI looks almost like magic.

Behind the scenes, however, gigantic physical infrastructure is running, including

● GPUs

● Semiconductor factories

● Data centers

● Power plants

● Cooling equipment

● Communication networks

● Water

● Rare metals

In other words, AI is a brain industry and, at the same time, a gigantic resource-consuming industry.

■ AI Is a Device That Turns Electricity into Heat

AI performs massive amounts of computation.

Massive computation requires a massive number of GPUs.

GPUs use large amounts of electricity.

And electricity almost always ends up as heat. (to be explained in detail in later chapters)

In other words, AI can also be described as a device that converts electricity into intelligence and heat.

■ The Bottleneck of the Future Is Cooling

What matters here is the point that if heat cannot be released, performance cannot be raised.

When GPUs get hot, they suffer

● Reduced performance

● Degradation

● Malfunctions

● Failure

In other words, in the future, what may matter more than how smart an AI you can build

is how much enormous heat you can process.

■ Data Centers Are Moving Closer to Power Plants

Today, giant data centers are being built around the world.

From now on, however,

● Power generation capacity

● Cooling capacity

● Water resources

may become more important.

In other words, in the future, what will matter is not where to run AI,

but where it can be cooled.

■ Cold Countries May Have the Advantage

In fact, in recent years, data center construction has been increasing in places such as

● Northern Europe

● Canada

● Cold regions

The reason is simple: they are easy to cool.

In other words, in the future, it may not be countries that can build AI

but countries that can cool AI that become strong.

Now, thinking about it that way, where does Japan stand?

■ Water Resources Become a Matter of National Competition

Cooling often uses large amounts of water.

In other words, from now on,

● Water

● Power generation

● Energy

● Cooling capacity

may themselves be directly linked to national competitiveness.

This is a very big change.

■ Semiconductor Competition Is Also Becoming a Heat Competition

The same goes for semiconductors.

In the AI era,

● Higher density

● Stacking

● Miniaturization

are accelerating.

As a result, however, escape routes for heat are beginning to disappear.

In other words, future semiconductor competition will be decided not only by how far you can scale down,

but also by how well heat can be made to work.

■ EV Competition Is Also a Heat Competition

In EVs, too,

● Fast charging

● Higher output

● Longer range

are advancing.

Behind the scenes, however, there is the enormous risk of thermal runaway.

In other words, future EV competition will be a contest not only of how high-performance they are,

but of how safely heat can be controlled.

■ The Robot Society Will Also Run into Heat Problems

In future robots,

● Compact, high-power designs

● Onboard AI

● Long operating hours

will advance.

As a result, heat density will rise sharply.

In other words, future robot competition will also become a matter not only of how smartly they move,

but of how well heat can be made to work.

■ An Era in Which AI Reveals Designs That Will Break

Furthermore, AI and sensors will make it possible to analyze enormous amounts of data, such as

● Temperature history

● Degradation

● Vibration

● Current

● Failure trends

As a result, in the future, it may become possible to see with extremely high precision not only why it broke,

but even which structures are prone to breaking.

In other words, in the future, thinking you are doing thermal management will no longer be enough.

■ Designs That Do Not Really Work Will Be Exposed

For example,

● Heat concentration

● Poor contact

● Insufficient heat dissipation

● Material degradation

● Tiny misalignment

● Differences in thermal expansion

Problems that were vague until now

may come to be identified by AI as structural problems.

To put it in extreme terms, the design that will break itself will be visualized.

As a result, even an increase in recalls may occur.

■ The Researcher's Perspective

Thermal control technology will become a national strategic technology

What will matter going forward includes

● Liquid cooling

● Radiative cooling

● Next-generation TIM

● High thermal conductivity materials

● Waste heat utilization

● Thermal simulation

● Power generation efficiency

In other words, thermal control may change from a component technology

into a national strategic technology.

■ The Shop-Floor Perspective

In the end, it stops at making it work in volume production

On the shop floor, however, very real problems occur, such as

● Contact

● Lifting

● Steps

● Thermal expansion

● Tiny misalignment

● Unstable automation

● Long-term degradation

In other words, in the future, rather than being “high-performance in theory,”

“working in volume production” will become the harder part.

■ What Matters from OTIS's Perspective

From OTIS’s perspective, what matters in future industries is not only high-performance materials.

What matters is

● Microfabrication

● Thin-film processing

● High-precision lamination

● Contact between dissimilar materials

● Automated supply

● Contact stability

● Volume production stability

in other words, processes that make functions work in volume production.

The more advanced the industry, the more important the final 0.1mm of heat becomes.

■ What OTIS Can Do

At OTIS, through

● Microfabrication

● High-precision die-cutting

● High-precision lamination

● Processing of thermal management materials

● Stacking of irregular shapes

● Support for automated supply

we may be able to contribute to making heat solutions work in volume production.

■ Outside OTIS's Expertise

On the other hand, OTIS is not a company that specializes in

● Power generation

● GPU design

● Semiconductor development

● National infrastructure design

● Data center operation

However, in the area of making future industries work in volume production processes, we may be able to play an important role.

■ Summary

In the future, the countries that can process heat may become strong

The AI race looks like a competition of intelligence.

Behind the scenes, however, an extremely physical competition has begun over

● Electricity

● Water

● Power generation

● Cooling

● Resources

● Thermal control

In other words, the future may become an era decided not only by how smart an AI you can build,

but by how much enormous heat can be made to work and processed.

And behind the scenes, technologies that support the final 0.1mm of heat will become increasingly important.

 

Column Supervisor: Koji Kakumoto (President, OTIS Co., Ltd.)

After studying abroad and working in planning and development at a trading company, he joined OTIS Co., Ltd. in 2011. While primarily working in the Corporate Planning Department, he has also served concurrently in manufacturing and technical divisions, and since 2018 has served as President, working to drive business growth and strengthen the organization.

 

This article is a general technical explanation intended for educational purposes and does not refer to any specific company, product, or technology.

Latest Articles

Related Articles

  • 0-2. Data Centers Will Become Cooling Plants ― In the AI Era, the World Begins to Use More Power for Heat Than for Computation ―

    0-2. Data Centers Will Become Cooling Plants ― In the AI Era, the World Begins to Use More Power for Heat Than for Computation ―

    With the evolution of AI, data center construction... Read More
  • 10-10. Proposals Based on the Customer’s Process (Part 3) ― A Component Only Becomes Good When One Looks Beyond the Drawing to “How It Is Used” ―

    10-10. Proposals Based on the Customer’s Process (Part 3) ― A Component Only Becomes Good When One Looks Beyond the Drawing to “How It Is Used” ―

    In Part 1, we looked at the perspective of designi... Read More
  • 10-10. Proposals Based on the Customer’s Process (Part 2) ― A Component Only Becomes Good When One Looks Beyond the Drawing to “How It Is Used” ―

    10-10. Proposals Based on the Customer’s Process (Part 2) ― A Component Only Becomes Good When One Looks Beyond the Drawing to “How It Is Used” ―

    In Part 1, we looked at the idea of understanding ... Read More

CONTACT

Okayama

+81-867-42-3690

Tokyo

+813-6810-4830

Contact