Redirection, Ownership & the Weapon Problem

Moving it to own it

Article II Forbids Owning a Celestial Body. Article VIII Says You Own a Space Object.

Nothing in between defines when a rock becomes a spacecraft.

The Outer Space Treaty of 1967 says two things that sit oddly together. Article II holds that outer space and celestial bodies are “not subject to national appropriation by claim of sovereignty, by means of use or occupation, or by any other means.” Article VIII holds that a state retains jurisdiction and control over any object it has launched into space, and that ownership of such objects is unaffected by their being in space.

An asteroid sitting in its own orbit is plainly the first thing. A spacecraft is plainly the second. The question nobody has answered is what a body becomes once a thruster is bolted to it and its orbit is under active control — whether it remains a celestial body that cannot be owned, or becomes a component of a space object that is.

The treaty text does not help. The 1972 Liability Convention defines a space object to include “component parts of a space object as well as its launch vehicle and parts thereof,” which is circular. The phrase “launched into outer space” suggests an asteroid does not qualify, since nobody launched it. But a captured body under continuous control, enclosed in a bag or fitted with a propulsion module, is hard to distinguish in function from a large satellite that happens to have been assembled in place.

So there is a reading — contested, untested, and not obviously wrong — in which moving an asteroid is the mechanism by which it becomes property, via the same instrument that forbids claiming it where it lies. No court has ruled. No state has tried it. The first commercial redirection will be the test case, and it will happen before the law is written rather than after. Observed

What a body is worth as a weapon

A Hundred-Meter Rock Arrives With the Energy of the Largest Bomb Ever Built

Kinetic energy scales with the cube of diameter, so the class of body matters more than anything else.

Impact energy is ½mv², and for a body of diameter d and density ρ the mass goes as d³. At a typical Earth-crossing encounter speed of 15 km/s:

For reference: Tsar Bomba, the largest device ever detonated, yielded about 50 Mt in October 1961. The Tunguska event in 1908 was on the order of 10 to 15 Mt from a body of roughly 50 to 60 meters. Density matters by nearly a factor of two, and asteroid densities are inferred rather than measured for almost every catalogued object — see the metal bodies page.

One line in that table is the one that matters, and it is not the largest. A hundred-meter body is comparable to the largest nuclear device ever built, and a three-hundred-meter body is nineteen times larger than anything anyone has ever detonated. The kilometer class is a different argument entirely, and as the next section shows, it is also out of reach.

900 kg and a decade

Aiming a 1.3-Billion-Kilogram Body Takes 0.084 Newtons

Give the problem ten years and the force required is about the weight of a AA battery.

The arithmetic that makes this a policy problem rather than a science-fiction one runs in three steps, and all three are elementary.

How much velocity change. To shift a body’s arrival by one Earth radius over a lead time T, the along-track velocity change needed is roughly one Earth radius divided by T. Over ten years:

6,371 km ÷ 10 yr = 20.2 mm/s

How much propellant. A 100-meter stony rubble pile masses about 1.31 × 10⁹ kg. Momentum required is mass times velocity change; divide by the exhaust velocity of a gridded ion thruster at 3,000 s specific impulse:

1.31×10⁹ kg × 0.0202 m/s ÷ 29,430 m/s = 898 kg of xenon

How much thruster. Spread that momentum over ten years and the thrust is 0.084 newtons, which at 3,000 s needs about 2.1 kilowatts of electrical power.

Dawn carried 416 kg of xenon and delivered 11.5 km/s of ion delta-v to itself over seven and a half years, so 898 kg is roughly two Dawns' worth of propellant — flown hardware, not a projection. The kilometer row is the useful bound: the bodies anyone can afford to move are the city-to-continent killers, not the global ones.

The reason the numbers are so small is that nothing here is in a hurry. A decade of continuous thrust at a tenth of a newton delivers the same momentum as a violent event, and an ion engine running for ten years is exactly what a commercial redirection tug is. The weapon and the tug are not similar. They are the same vehicle. Modeled

Nothing on any list

Every Other Weapon of This Scale Needs Something Controlled. This One Needs Xenon.

The non-proliferation architecture of the last eighty years has no purchase here.

Arms control works by controlling inputs. Nuclear weapons need fissile material, which requires enrichment or reprocessing, which requires facilities that are large, detectable and safeguarded. Chemical and biological weapons need precursors and equipment covered by the Australia Group and the relevant conventions. Ballistic delivery is constrained by the Missile Technology Control Regime. In every case there is a choke point, and the regime sits on it.

A redirection weapon has no choke point. It needs:

Xenon is an industrial gas. Hall and gridded ion thrusters fly on hundreds of commercial satellites. A 2 kW solar array is a smallsat bus. Deep-space navigation is a published discipline. Nothing in the list is restricted, and nothing about acquiring it looks like weapons development.

What it does need is something no treaty has ever had to control: a decade of uninterrupted, unremarkable operation. That is a real barrier — it requires an organization that survives ten years, a deep-space network, and a body already surveyed. It is not a barrier any existing regime is built to see, because it looks exactly like a business.

Trajectory is software

An Inspection Certifies Nothing That Survives the Next Command Load

The civil and military versions are the same hardware in the same configuration, differing only in a number.

Arms control also works by inspection. An inspector can look at a centrifuge cascade and determine what it is enriching to; at a reactor and determine whether fuel has been diverted; at a missile and count its stages. Each is a physical fact about an object, and it stays true after the inspector leaves.

There is no equivalent here. A tug pushing a body toward a processing station and a tug pushing the same body toward Earth are identical in every respect an inspector can examine: same thruster, same propellant load, same power system, same attachment, same mass. The difference lives entirely in the target state loaded into the guidance computer, and that can be changed by a command from the ground at any point in a decade-long burn.

This inverts the usual relationship between capability and intent. In every other domain, building the capability is the hard, visible, controllable step, and the decision to use it is instantaneous and unverifiable. Here the capability is commercial, visible and encouraged, and the decision to use it is the only controllable moment — but it is a software change that can be made at any point during the ten years the capability is being exercised for legitimate purposes.

The same finding appears, in a milder form, at every electromagnetic launcher in this regime: a mass driver that throws cargo can throw an impactor, and what separates the two is the aim point. Redirection is the same problem with a thousand times the payload and a decade of warning that nobody is watching for. Modeled

Default-miss costs ten times

The Only Real Mitigation Has a Price, and It Has to Be Paid Before There Is an Industry

Requiring a body to miss Earth if control is lost means committing late, and committing late means ten times the propellant.

The mitigation that keeps being proposed, and which is correct, is a default-miss requirement: any body under active redirection must be on a trajectory that misses Earth if all control ceases, with the final approach committed only by a late, small, verifiable correction. Loss of control then produces a miss rather than an impact, and the moment of commitment is compressed into a window short enough to observe.

It is not free, and the cost falls out of the same arithmetic as the threat. The velocity change needed scales inversely with lead time, so a correction committed one year out costs ten times what the same correction costs ten years out:

The same inverse relationship that makes the weapon cheap makes the safeguard expensive. A regime that permits ten-year commitment gets 898 kg of propellant and no verification window; a regime that demands one-year commitment gets a year to watch and a nine-tonne propellant bill. Verifiability and propellant mass trade directly against each other, which is why this is an architecture decision rather than a policy overlay.

Two supporting measures cost almost nothing and are worth naming alongside it: independent tracking, since a redirected body is an observable object and the Vera Rubin Observatory will be watching this population anyway, and a public catalogue of commanded versus ballistic trajectories, so that any divergence between what an operator says it is doing and what the body is actually doing is visible to third parties without cooperation.

What is demonstrated

Deliberate Orbit Change Stopped Being Theoretical in September 2022

DART cost about the price of a science mission and worked on the first attempt.

NASA’s DART spacecraft struck the moonlet Dimorphos in September 2022 at 6.1 km/s and shortened its orbital period around Didymos by roughly 33 minutes — far more than the mission’s own success criterion, because ejecta thrown off the surface carried away several times the spacecraft’s own momentum. It was a 570-kilogram vehicle built from standard components, and it was a first attempt. ESA’s Hera follows to survey the result in detail and to measure the mass and structure the momentum calculation depends on.

The planetary-defense framing is genuine and this site does not dispute it. What DART also established is that a small body’s orbit can be changed deliberately by an organization with a modest budget, and that the momentum enhancement from ejecta makes kinetic impact more effective than the naive calculation suggests — a multiplier that works identically whichever direction the body is being pushed.

Rendezvous capability is the other half and it is more relevant than impact. China’s Tianwen-2 matched orbits with the near-Earth object Kamoʻoalewa in July 2026, operated at 20 kilometers range, and is sampling a body roughly 18 to 20 meters across before returning its capsule to Earth in November 2027 and continuing to a main-belt comet. Matching orbits with a small body and working at it for months is precisely the hard part of attaching anything to one. Japan demonstrated the same class of operation twice with Hayabusa and Hayabusa2, and the United States with OSIRIS-REx. Four programmes have now proven the operational capability that redirection requires, none of them for that purpose. Observed

The commercial layer is thinner and moving the same direction. TransAstra’s stated approach is capture-and-relocate — enclosing a small body and moving it somewhere convenient — which is redirection by another name and entirely legitimate as a mining method. SpaceX named asteroid mining in its May 2026 IPO filing. No commercial redirection has been attempted, and the first one will establish more precedent than anyone intends.

Two legal questions

Who Owns the Ore Is Mostly Settled. Who Owns the Rock Is Not.

National law has answered the easy question and left the strange one alone.

There are two distinct property questions in this regime and they are usually run together. The first — can you keep what you extract? — has been answered in practice, if not in principle. The United States Commercial Space Launch Competitiveness Act of 2015 grants US citizens rights to resources they recover; Luxembourg followed in 2017, then the United Arab Emirates and Japan. The Artemis Accords, from 2020 onward, state that resource extraction does not constitute national appropriation under Article II, and dozens of states have signed.

That is a workable answer built on an analogy to fishing: the ocean is nobody’s, the catch is yours. It is also not universally accepted. Russia and China are not Accords signatories, and a body of scholarship holds that the Accords interpret Article II rather than apply it. The 1979 Moon Agreement, which declares celestial resources the common heritage of mankind and requires an international regime to govern exploitation, was ratified by too few states to matter but has never been repudiated either.

The second question — can you own the body itself by moving it? — has no answer at all, and it is the one this page is about. The fishing analogy breaks down, because relocating a body is not harvesting from a commons; it is taking the commons somewhere else. Nothing in the Accords, in national law, or in the treaty addresses it.

None of this resolves in a page, and it is not this site’s job to resolve it. What the arithmetic establishes is narrower and harder to argue with: a capability that needs no controlled input, cannot be distinguished from commerce by inspection, and commits a decade before it acts is outside the reach of every verification method the last eighty years produced. The treaty was drafted in 1967, when no state could have built one. The first entity that can will not be a state.

Record

Provenance and Data Sources