AI consumes electricity.

Thinking about Japan's next industrial map from Akita's '2 trillion yen AI data center project'

The large-scale AI data center project in Akita was featured on the front page of yesterday's Nihon Keizai Shimbun.

It is a maximum 500MW scale project with construction costs of approximately 2 trillion yen.

Aiming for operation in the early 2030s, it is said that collaboration with overseas investors is also being considered.

Of course, this is still at the planning stage.

Even so, I could not read this news simply as a story about 'building a huge data center'.

When you are involved in the energy business, you can see the major industrial structural changes that might happen in Japan in the future on the other side of this news.

What is powering AI?

When it comes to generative AI, in many cases,

semiconductors, GPUs, large language models, data, and software are mentioned.

These are the words that are lined up.

However, no matter how excellent an AI is, it will not run for even a second without electricity.

And the more advanced AI becomes, the more the amount of computation increases, and the power demand to run it also grows.

That is why I

believe that the AI industry is also a massive power-consuming industry.

I think of it that way.

Thinking about it that way, what is important in this news is not just the '2 trillion yen' figure.

Rather,

'Why Akita?'

might be a more important question.

Is industry heading toward 'places with electricity'?

Akita is a region with great potential for renewable energy, including offshore wind power.

Until now, the common idea was to transport electricity generated in rural areas to urban centers.

However, if AI data centers that consume massive amounts of electricity start locating in rural areas, this structure will change.

Instead of transporting electricity to cities,
bring industry to where the electricity is.

There is a possibility that we are entering such an era.

Looking back, the locational requirements for industry have changed with the times.

There was an era when ports were necessary.

There was an era when railways were necessary.

There was also an era when logistics hubs and factories gathered around highway interchanges.

And in the AI era,

regions that can secure large and stable amounts of electricity

might become the new condition for industrial location.

If that happens, the meaning of the term 'regional revitalization' will also change.

It is not just about attracting population.

It is not just about increasing tourists.

Starting from the energy resources that exist in the region,

generate,
store,
calculate,
create industry,
and generate employment.

There is a possibility that such regional revitalization will become a reality.

However, AI cannot be powered by renewable energy alone.

There is something we must not forget here.

The power generation capacity of renewable energy sources like solar and wind varies depending on natural conditions.

On the other hand, data centers operate 24 hours a day, 365 days a year.

In other words,

"Having abundant renewable energy" and "being able to supply stable power to AI data centers" are not synonymous.

Something is needed to bridge that gap.

Storage batteries.

Transmission and distribution networks.

Supply and demand adjustment.

Backup power sources.

Energy management.

And power procurement itself.

I am currently involved in storage batteries and energy-related businesses, and there is something I strongly feel since entering this world.

"Being able to generate power" and "being able to deliver usable power" are different.

Delivering the necessary amount, at the necessary time, with the necessary quality.

Only when that is achieved does electricity become an industrial infrastructure.

That is why I believe,

the AI revolution is simultaneously a power and energy storage infrastructure revolution.

I think so.

It is precisely because we are in the age of AI that the value of the real world is increasing.

It may seem like a slightly different topic, but yesterday I was talking about closing at a gathering of business owners.

As AI advances, much of sales activity will become automated.

Prospect identification.

Information gathering.

Drafting text.

Customer analysis.

We will be able to rely on the power of AI for these tasks more and more in the future.

However, in the end,

"I will buy from this person"

"I will receive this treatment"

"I will work with this company"

When making these decisions, human interaction and trust remain.

And reading an article about AI data centers, I felt a strange commonality.

The more AI evolves,

the more "people" become important in sales,

and the more "electricity" becomes important for AI infrastructure.

Neither can be completed by digital means alone.

The "visible foundation" that supports the "invisible"

We are captivated by AI, an invisible intelligence.

However, the places that create that intelligence have massive buildings, servers, power lines, electrical equipment, and people.

Ultimately, it is a very physical world.

That is precisely why,

"How should Japan utilize AI?"

I believe that question alone is not enough.

What is needed from here on out is,

Where in Japan will we build the infrastructure to power AI?

How will we secure that electricity?

Who will invest in it?

What kind of industries and jobs will we leave for the region?

I think these are the questions we should be asking.

This project currently being considered in Akita.

It is still a plan for the future.

Even so, I do not see it merely as an article about a large-scale investment,

but as news that indicates the possibility of Japan's next industrial map being redrawn.

I am watching it as such.

AI consumes electricity.

And where, how, and by what means will we generate, store, and deliver that electricity?

Perhaps the key hints for considering Japan's next growth industry are hidden right there.