In September 2022 a NASA spacecraft deliberately crashed into a harmless asteroid 11 million kilometres away and shifted its orbit by about 32 minutes — the first time humans had ever moved another world, the moment planetary defence stopped being theory and became something we have actually done.

On 26 September 2022, NASA’s Double Asteroid Redirection Test spacecraft struck Dimorphos, a roughly 150-metre moonlet orbiting the larger asteroid Didymos, about 11 million kilometres from Earth. The roughly 600-kilogram spacecraft hit at about 6 kilometres per second and shortened the moonlet’s orbit around Didymos. It was the first time anyone had deliberately changed the orbit of an asteroid system in a planetary-defence test, and it remains the only successful test of its kind.

That is a real first, and worth stating without inflation. Dimorphos is a small moonlet, not a planet, and it stayed bound to its companion throughout. Neither asteroid was ever on a course to hit Earth. The system was chosen precisely because it was safe and because the change could be measured from the ground.

What the numbers actually say

NASA’s minimum threshold for calling the test a success was a period change of 73 seconds. The mission cleared it by a wide margin. The orbital period was first reported as having shortened by about 32 minutes, and the peer-reviewed figure, published by Cristina Thomas and colleagues in Nature in 2023, settled at 33.0 minutes, give or take a minute.

The size of that change is the interesting part. If DART had simply transferred its own momentum to Dimorphos in a clean collision, the expected period change was around seven minutes. The measured result was several times larger. The difference came from recoil. When the spacecraft hit, it blew a large plume of rock and dust off the surface, and the momentum of that escaping material pushed back on the asteroid like exhaust from a thruster.

Researchers quantify this with a factor called beta, the ratio of total momentum delivered to the momentum the spacecraft carried in. A separate analysis led by Andrew Cheng put beta at around 3.6, meaning the ejecta did most of the work. By that measure the impact was far more effective than the bare collision would have been.

Why that efficiency may not transfer

The recoil that made DART so effective is also the part that is hardest to count on. Beta depends on what the target is made of. A loose rubble pile throws off a great deal of debris and recoils hard. A denser, more solid, or differently structured asteroid would shed less, and the same spacecraft might move it far less. Estimates of Dimorphos’s own beta still range from about 2.2 to 4.9, because its mass is not precisely known.

The aftermath was also messier than a single clean nudge. Dimorphos is a weak, boulder-strewn body, and the impact appears to have reshaped it rather than just leaving a crater. Its orbit kept evolving for weeks afterward and picked up a slight eccentricity, and follow-up studies have raised the possibility that its rotation was knocked into a wobble. None of this undoes the result. It does mean the clean headline figure sits on top of a more complicated physical event.

The part DART did not test

A kinetic impact changes an asteroid’s speed by a tiny amount. DART altered Dimorphos’s orbital velocity by a few millimetres per second. That works as planetary defence only because a small change applied years or decades before a predicted collision compounds into a large shift in position by the time the asteroid reaches Earth’s neighbourhood. The technique is useless without warning, which means the binding constraint is detection, not the impact itself. Many of the asteroids capable of regional damage are exactly the class we have not yet catalogued well enough. NASA’s own planetary-defence work estimates that around 25,000 near-Earth objects larger than 140 metres exist, and fewer than half have so far been found and tracked.

So DART demonstrated that the method can work on a body of this kind. It did not demonstrate that we could reliably stop a specific incoming asteroid, which would depend on finding it early enough and knowing enough about it to predict how it would respond.

The other half of the experiment

DART was always designed as one half of a two-part programme. The European Space Agency’s Hera spacecraft is the second half, built to return to Didymos and measure what ground telescopes cannot. Hera launched in October 2024, swung past Mars in March 2025, and after a deep-space manoeuvre in February 2026 is on track to reach the asteroid system in late 2026, with ESA targeting November, a month ahead of the original December plan.

Its central task is to weigh Dimorphos. The momentum-enhancement figure that made DART look so effective carries roughly a 10 per cent uncertainty, almost all of it from the unknown mass. Hera should greatly reduce that uncertainty by measuring Dimorphos’s mass directly. It will also survey the crater and document the reshaping. Only then does the 2022 result become a calibrated number rather than a strong but loosely constrained one.

The thing to watch is Hera’s arrival. Until it delivers its measurements, DART stands as a clear demonstration that a kinetic impactor can move a small asteroid moonlet, and an incomplete account of exactly how efficiently it did so. The late-2026 rendezvous is what begins turning the first result into the second.