Why I chose water in the age of AI
I have spent most of my career working in IT. I watched the internet spread, smartphones arrive and, now, AI begin to change society at extraordinary speed.
I love new technology. I have seen many moments when technology made something possible that had never been possible before. That is why I see enormous potential in AI.
Yet as I became involved with AI, one thought would not leave me:
AI needs water.
The water behind AI
When we ask AI a question, an answer comes back in seconds. Looking only at the screen, it is difficult to sense the natural resources being used. Behind that screen, however, are enormous data centers. Large numbers of computers operate there, consuming electricity and water to keep cool.
Manufacturing the semiconductors and GPUs those computers depend on also requires vast amounts of water — enough to make beverage manufacturers seem small by comparison.
While researching the relationship between AI and water, I came across a study estimating that around 20 exchanges with ChatGPT could use roughly 500 ml of water. Of course, that figure varies greatly with location, equipment, climate and the calculation method.
Even so, it shocked me. There have also been data centers in North America that stopped operating when water could no longer circulate.
The technology we use so casually every day and the water beneath our forests are connected by a single line. That is how I began to see it.
Water does not disappear, but it does not return immediately
Water circulates around the Earth. Rain falls, soaks into forests and soil, enters the ground, becomes rivers, flows into the sea, evaporates and falls again. Water is a circulating resource.
But the water we use today will not necessarily return to the same place, in the same condition, tomorrow.
This is particularly true of groundwater. Rain and snowmelt seep into the ground, move slowly through soil and rock, and are stored as cool groundwater. That takes a long time.
The speed at which people use water and the speed at which nature replenishes it: as AI grows, I have come to believe that we must think carefully about both.
Japan has a wealth of forests
Around 70% of Japan’s land is forested. I see that as a tremendous asset.
But the presence of a forest alone does not mean it will nurture water. Many forests were planted by people and later received too little care. Trees become crowded, light no longer reaches the ground and undergrowth stops growing. Rain strikes the ground directly, and the soil becomes compacted.
Then rain and snowmelt find it harder to soak into the ground. Water that should have entered forest soil and slowly traveled underground instead runs over the surface.
I saw an opportunity there.
Begin by observing water
What amenowa first set out to do is actually quite modest:
Observe groundwater.
How does groundwater respond when rain falls? When snow accumulates and then melts in spring, how long does it take to reach the groundwater? If we change the condition of a forest, does the water’s movement change?
First, measure. Then, understand. After that, manage the forest and create conditions that allow rain and snowmelt to soak into the ground more easily.
Rather than the technical term “groundwater recharge,” I like to call this “nurturing water.”
I do not want to stop technology
I did not start amenowa because I oppose AI or data centers. Quite the opposite: I believe in the future of technology. I think AI will transform our lives.
That is precisely why I want to think just as seriously about the water and forests that support that future.
We use technology. We use energy. We use water. Then could we become people who nurture water, as well as people who use it?
That is amenowa’s starting point.
A project far larger than ourselves
The vision is enormous. Honestly, we cannot achieve it alone.
Forests have their specialists, and groundwater has its specialists. People who have lived in a place know things that no one else can know. Local government, universities, companies and residents all need to be involved.
So we are not trying to do something huge from the outset. We begin by observing groundwater in one place. We look at rain, snow and forests. Gradually, we learn how water is replenished and where it goes. From there, we nurture it.
Even in a future where AI is part of everyday life, I want children to be able to turn on a tap and find cool water flowing as naturally as ever. Rivers should keep flowing, and forests should keep storing water. I want to leave that ordinary abundance to the next generation and the generations after it.
Water is not simply a resource to use. It is something we can nurture ourselves.
Three years ago, I started amenowa to put that idea into practice. Together with Rusutsu Resort in Hokkaido, we began a project to visualize the water cycle and monitor the natural environment.
Our small project has begun attracting attention from many people. Our challenge has only just begun.