Near-Earth Objects

What they are

Near-Earth Objects: Belt Material on Orbits That Come to Us

Perihelion inside 1.3 AU, four dynamical groups, and a residence time far shorter than the age of the solar system.

A near-Earth object is any small body whose closest approach to the Sun falls within 1.3 AU. They divide into four groups by orbit: Atiras stay entirely inside Earth’s orbit, Atens have smaller orbits than Earth’s but cross it, Apollos have larger orbits and cross it, and Amors approach without crossing. Roughly 40,000 are catalogued.

None of them originated where they are. Gravitational resonances with Jupiter — the Kirkwood gaps — pump the eccentricity of main-belt fragments until they fall inward, and the Yarkovsky effect, an asymmetric thermal push from re-radiated sunlight, slowly drifts smaller bodies into those resonances. The population turns over on a timescale of millions of years, short against the age of the solar system, which means it is continuously resupplied from the belt.

For an industrial reading that origin is the point. Near-Earth objects are belt material delivered to the doorstep, and anything learned about processing them applies to the parent population. They are a sample set and a proving ground rather than a separate destination. Observed

Cheaper than the Moon

Some Near-Earth Asteroids Cost 3.8 km/s from Low Earth Orbit

Less than the lunar surface, and less than geostationary transfer plus circularization.

An object already on an orbit close to Earth’s needs only a small nudge to match. The most accessible known targets sit near 3.8 km/s from low Earth orbit for a rendezvous, against 5.9 km/s to land on the Moon. No gravity well to climb out of at the far end, no descent propulsion, no landing gear in the usual sense — arriving is a matter of matching velocity and station-keeping.

The 3.8 km/s figure applies to a handful of favorably placed objects out of tens of thousands, and to rendezvous rather than to sample return, which costs again on the way home. Most catalogued near-Earth objects are considerably more expensive than the Moon.

That number is what every asteroid mining pitch deck opens with, and taken alone it is true. The problem is what comes attached to it.

The cadence trap

The Closer the Orbit, the Longer the Wait Between Launch Windows

The same synodic formula that favors the belt turns against near-Earth objects, hard.

Launch opportunities arrive at the synodic period, set by the difference in angular rates:

1 / S = | 1/T_Earth − 1/T_target |

Cheap access requires an orbit similar to Earth’s. An orbit similar to Earth’s has a period similar to Earth’s. And when the two periods are nearly equal, the difference in the denominator goes to zero and the synodic period goes to infinity. The property that makes a target cheap is the same property that makes it rare.

Eccentricity and inclination complicate individual cases, and a mission can accept a more expensive transfer to fly outside the ideal window. The structural relationship holds regardless: within the near-Earth population, delta-v and cadence trade directly against each other.

So the belt’s cadence inversion has a mirror image here. Ceres gets 7.8 windows a decade because its orbit is nothing like Earth’s. The cheapest near-Earth object might offer one opportunity in fourteen years. A program cannot iterate on a target it reaches once per decade. Observed

Iterating across a population

Why the First Wave of Asteroid Mining Companies Failed

Reliability is bought with repetition, and a decadal window does not sell any.

Every reliable space system got that way by flying repeatedly and fixing what broke. Falcon 9 flew more than four hundred times before anyone called it routine. Lunar programs get a window every few days and Mars programs every 26 months, which is slow but survivable — a failure in one window is a lesson applied in the next.

Near-Earth targets break that loop. A failure at a body reachable once per decade is not a lesson applied next time; there is no next time within the life of the company. The only available form of learning is across the population rather than at any single body — fly many cheap, near-identical spacecraft at many different targets and learn from the aggregate.

That is exactly the difference between the two commercial waves. Planetary Resources and Deep Space Industries, both founded in the early 2010s with heavyweight backing, planned flagship prospecting architectures and were absorbed in 2018 and 2019 without reaching an asteroid. The current wave designs around replication instead: AstroForge builds small, cheap, repeatable spacecraft and accepts losses, which is the correct structural response even though its first vehicle, Odin, was lost to a solar array failure shortly after launch in February 2025. Karman+ is pursuing carbonaceous water targets with a demonstration aimed at February 2027. TransAstra works on capture-and-relocate hardware rather than in-place processing. The commercial flight record to date is 0 for 2.

One consequence deserves stating plainly, because it undercuts the usual framing. A near-Earth object is not a mine, because a mine is a place you return to. It is a single harvest on a schedule set by orbital mechanics, and the business that works is the one that treats targets as disposable and the spacecraft as the product. Modeled

Four kinds, not one

Near-Earth Objects Are Not One Material, and Dead Comets Are Among Them

A target list that mixes the types produces hardware designed for an average body that does not exist.

The population spans at least four compositional classes, and the machine that works on one will not work on another. Carbonaceous bodies like Bennu and Ryugu carry hydrated minerals, organics and water, and their material is dark, porous and so fragile that returned grains break under gentle handling. Silicate bodies like Itokawa and Eros are ordinary rock and gravel. Metal-rich bodies carry nickel-iron, and no sample has ever been returned from one.

The fourth class is the one that is routinely left off target lists. Some near-Earth objects are dormant or extinct cometary nuclei — bodies that have lost their surface volatiles and now look like asteroids while retaining buried ice and an entirely different internal structure. 3200 Phaethon, the parent of the Geminid meteor shower and the target of JAXA’s DESTINY+ mission, sheds material at perihelion without a normal coma. 2P/Encke is the classic transitional case.

For a water-seeking operation these are potentially the best targets in the population, and for an anchoring or drilling system they are the worst-characterized. Sorting the catalog by accessibility alone, as most prospecting lists do, sets the wrong mental model: it implies a fleet of similar bodies differing mainly in how far away they are. Type belongs on the axis alongside delta-v, and the catalog does not currently carry it. Projected

Kamoʻoalewa shows how exotic the population can get. This quasi-satellite of Earth, which China’s Tianwen-2 reached in July 2026, has a spectrum matching lunar samples rather than any asteroid class — it appears to be a fragment of the Moon, blasted off by an impact and now orbiting alongside us. It also turned out to be roughly 18 to 20 meters across, about half the size ground-based observation had estimated.

Who has been

Every Asteroid Sample Ever Returned Came from a Near-Earth Object

About 127 grams in total, from three missions, all of it from bodies that came to us.

NEAR Shoemaker touched down on Eros in February 2001, the first landing on an asteroid and one not designed for it. Hayabusa brought back grains from Itokawa in 2010. Hayabusa2 returned 5.4 grams from Ryugu in December 2020 and continued on an extended mission, flying past the contact binary 98943 Torifune in mid-2026. OSIRIS-REx returned 121.6 grams from Bennu in September 2023 and, as OSIRIS-APEX, is now headed for Apophis.

DART struck Dimorphos in September 2022 and shortened its orbital period around Didymos by about 32 minutes — the first demonstration that a small body’s trajectory can be changed deliberately. ESA’s Hera arrives at the same system to survey the result in detail. The implications of that capability run well past planetary defense, and they are handled on Redirection, Ownership & the Weapon Problem.

Tianwen-2 is the current mission and the most ambitious sequence anyone has flown here: it arrived at Kamoʻoalewa on 2 July 2026 at 20 kilometers range, samples 200 to 1,000 grams through April 2027, releases its return capsule at an Earth flyby in November 2027, and then departs for the main-belt comet 311P/PANSTARRS with arrival in January 2035. One spacecraft, a near-Earth sample return and a main-belt rendezvous in a single mission.

The largest near-term event needs no spacecraft at all. Apophis passes within roughly 32,000 kilometers of Earth on 13 April 2029 — inside geostationary altitude, visible to the naked eye from parts of Europe and Africa. OSIRIS-APEX will be there, and ESA has proposed Ramses to rendezvous before the encounter so the body can be observed while Earth’s tides work on it. Observed

The catalog problem

The Instruments That Matter Next Measure Size and Composition, Not Position

Rubin finds them. NEO Surveyor says how big they are.

The Vera Rubin Observatory began its survey on 30 June 2026 and added more than 11,000 asteroids from six weeks of pre-survey data alone. Projections put the eventual near-Earth yield on the order of 127,000 objects, lifting completeness for bodies 140 meters and larger from roughly 40% to 70%. The population a prospector chooses from is about to grow several-fold.

What Rubin measures is position and brightness, and brightness conflates size with albedo — a small bright body and a large dark one look alike. That is the error that made Kamoʻoalewa look twice its true size and that makes every remote composition call provisional. NEO Surveyor, launching in 2027, observes in the infrared and therefore returns size and albedo directly, which is the measurement any target-selection process actually needs. It is the instrument that turns a catalog of dots into a catalog of bodies.

Until it flies, every commercial target list is built on inference of the kind that has now failed twice in a row on the two cases anyone checked closely — Psyche’s density and Kamoʻoalewa’s size. Prospecting is the load-bearing activity in this regime and it has barely started.

Record

Provenance and Data Sources