For most of human history, the Moon was something we could only look at.

Then, in the twentieth century, we sent spacecraft there. We photographed it, mapped it, landed robotic probes on it and, eventually, walked on its surface.

Now another change is beginning.

Artificial intelligence is becoming part of lunar exploration.

This does not mean that an AI robot is about to wander around the Moon having conversations with itself. The reality is much more interesting.

AI is being developed to help scientists study enormous amounts of lunar data, recognise geological features, search for water ice, guide robotic explorers and eventually help spacecraft and rovers make some decisions for themselves when they are temporarily out of contact with Earth.

One of the newest developments arrived in September 2026. NASA and IBM released a new artificial intelligence system designed specifically to study the Moon.

Teaching AI to understand the Moon

On 10 September 2026, NASA and IBM announced the open-source NASA-IBM Lunar Foundation Model.

In simple terms, it is an AI model trained to recognise and interpret features of the lunar environment.

Scientists have already accumulated an enormous amount of information about the Moon. NASA’s Lunar Reconnaissance Orbiter, which has been orbiting the Moon since 2009, has photographed the surface in extraordinary detail. Other missions have measured temperature, gravity, mineral composition, surface structure and other properties.

The new AI model brings information from many of these sources together. IBM says the associated dataset contains more than 30 aligned layers of information derived from instruments aboard four different missions, including NASA’s Lunar Reconnaissance Orbiter and GRAIL missions, together with data from Japan’s SELENE/Kaguya mission.

That is important because a human scientist looking at one photograph can see a great deal.

But imagine trying to compare tens of thousands of photographs with temperature measurements, mineral maps, elevation data, radar observations and gravity measurements covering the entire Moon.

That is where AI becomes useful.

It can search through huge quantities of information and help identify patterns that deserve a scientist’s attention.

The intention is not to replace lunar scientists. It is to give them a new tool.

Looking for craters, volcanoes and water

NASA and IBM have initially identified some particularly interesting jobs for the lunar AI.

One is finding and mapping small craters that have never been formally catalogued. Another is investigating the Moon’s volcanic history.

But perhaps the most intriguing task is looking for evidence of ice near the lunar poles. IBM says the model has been tested for identifying features including possible ice deposits, craters and volcanic formations.

That brings us to one of the biggest stories in modern lunar exploration.

Water.

There is water on the Moon

For generations the Moon was generally imagined as an almost completely dry world. We now know the story is more complicated.

Water ice has been detected in some of the Moon’s permanently shadowed polar regions, including around the south pole.

These locations are extraordinary places.

Because the Moon’s rotational axis is only slightly tilted, sunlight reaches parts of the polar landscape at an extremely shallow angle. Some deep crater floors therefore never see the Sun at all.

Temperatures inside them can become very low, allowing water ice and other volatile substances to survive for extremely long periods. NASA’s Lunar Reconnaissance Orbiter has helped identify some of these exceptionally cold regions and evidence of water ice near the south pole.

The ice is scientifically interesting because it may preserve information about the early history of the Moon and Solar System.

But it may eventually be useful for something else.

Human beings.

Why lunar water could be so valuable

If future astronauts can obtain water from the Moon rather than transporting all of it from Earth, lunar exploration could become more practical.

Water can obviously be used for drinking. But water is made from hydrogen and oxygen.

Oxygen could potentially be produced for breathing. Hydrogen and oxygen can also be used as rocket propellants.

So a deposit of lunar ice could eventually become much more than something interesting for scientists to examine. It could become a resource.

That is why there is so much interest in the Moon’s south polar region.

Finding that ice is not simply a matter of landing beside a sparkling frozen lake. The water may be mixed into lunar soil, buried below the surface or concentrated in places that are extremely difficult to reach.

This is precisely the sort of search in which AI may become useful. It can compare different kinds of observations and help researchers work out where a rover, drill or future astronaut might have the best chance of finding something interesting.