The US Says It Has Weapons in Orbit. The Treaty Probably Allows It, and That Is the Problem

There is a particular kind of silence that follows a government saying something everybody already assumed. On 14 September, at the Air and Space Forces Association conference at National Harbor, the US Secretary of the Air Force, Troy Meink, said this:

The United States has on-orbit space control weapons capable of defending the Joint Force against hostile adversary action.

Troy Meink, Secretary of the Air Force, 14 September 2026

That is the whole disclosure. One sentence. No system named, no orbit given, no numbers, no date of deployment. Asked for detail, Meink said he was not going to talk about specifics, whether anything had been tested, or the type of weapon.

Within a day the Space Force had gone further rather than walking it back. A spokesperson confirmed the wording and then defined the mission set, describing space control as employing kinetic and non-kinetic means to affect adversary capabilities through disruption, degradation and, if necessary, destruction, available for offensive as well as defensive purposes at the direction of a combatant command.

Read those two statements together and the gap between them is the story. The announcement was dressed as defence. The doctrine it sits inside is not limited to defence.

I want to make a narrow argument here, because the wide one is being made badly almost everywhere else this week. The argument is this: nothing unlawful has been established, the law is the problem, and the people who have spent forty years failing to fix the law now have a much harder job than they did on 13 September.

What the treaty actually says

Article IV of the Outer Space Treaty has two paragraphs and they do very different things.

The first paragraph is the famous one. States undertake not to place in orbit around the Earth any objects carrying nuclear weapons or any other kinds of weapons of mass destruction, not to install such weapons on celestial bodies, and not to station such weapons in outer space in any other manner.

Notice what the sentence is keyed to. Every limb of it, including the catch-all at the end, attaches to nuclear weapons or other weapons of mass destruction. The catch-all widens the manner of placement. It does not widen the class of weapon. A conventional weapon in Earth orbit is not caught by paragraph one, and never has been.

The second paragraph is where the language people have been quoting all week actually lives. The Moon and other celestial bodies shall be used exclusively for peaceful purposes. Military bases, installations and fortifications are forbidden there, as is the testing of any type of weapons and the conduct of military manoeuvres.

That is a real prohibition, and it is a broad one. It applies to the Moon. It does not apply to low Earth orbit.

I have now read several respectable outlets describing the treaty as reserving space for peaceful purposes, full stop. National Public Radio called space under the treaty a shared global domain reserved exclusively for peaceful applications. That is simply not what the text says, and the error matters, because it lets readers conclude that a violation has occurred when the more uncomfortable truth is that no violation is needed for this to be alarming.

So is it lawful?

On the face of the treaty, yes, probably. The United States asserts compliance and the assertion is not obviously wrong.

There are three places where it could still get complicated, and none of them can be assessed on the information released.

Article III and the UN Charter

Article III requires that space activity be carried on in accordance with international law, including the UN Charter. That is the hook by which Article 2(4) on the threat or use of force, and Article 51 on self-defence, reach into orbit.

Placing a weapon is not using force. But a weapon whose purpose is announced as defending against hostile action, in a domain with no agreed rules of engagement and no agreed threshold for what counts as an attack, is a threat posture. Whether an announcement of capability crosses into a threat of force under Article 2(4) is genuinely contested, and one sentence with no stated conditions of use gives nobody enough to judge it.

Article IX and the duty to consult

Article IX requires a state that has reason to believe a planned activity would cause potentially harmful interference with the activities of others to undertake appropriate international consultations before proceeding.

Note the shape of that duty. It is triggered by planned activity, it is owed before proceeding, and it is a duty to consult rather than a duty to obtain consent. I could find no public record of any Article IX consultation preceding this deployment. Absence of reporting is not proof of absence, and I would not assert that none took place. But if one did, nobody has said so, and the provision is worth considerably more attention than it gets.

The peaceful purposes split

Underneath all of this sits an interpretive argument as old as the treaty. Does peaceful mean non-military, or does it mean non-aggressive?

State practice has answered this decisively in favour of non-aggressive. Orbit has been used for military reconnaissance since 1957 and nobody seriously argues that weather satellites for artillery or GPS for munitions guidance breach the treaty. The argument is not close on the law. It is only close on the politics.

The part that should worry you

Here is what I keep coming back to. The treaty is fine. The treaty is doing exactly what it was designed to do, which was to stop the two superpowers of 1966 putting nuclear weapons over each other’s cities. It has done that for fifty nine years across one hundred and eighteen states parties, and that is an extraordinary record.

What the treaty was never designed to do is govern the thing that has just been announced. And the machinery built to fill that gap has failed, repeatedly, in public, for decades.

Consider what the record looks like.

The Conference on Disarmament has had prevention of an arms race in outer space on its agenda since 1985. An ad hoc committee ran from 1985 to 1994 and achieved essentially nothing. The Stockholm International Peace Research Institute describes the discussions as deadlocked for decades, stuck on the two questions that matter most, which weapons to prohibit and how to verify.

The Russia and China draft treaty, the PPWT, has been on the table since 2008 and revised in 2014. The longstanding US objection is that it does not define a weapon in space, does not cover ground-based threats, and has no verification mechanism. Those objections are, in my view, substantially correct, which is a problem, because the alternative on offer is nothing.

At the Security Council in April 2024 a US and Japanese draft reaffirming the Article IV obligation was vetoed by Russia, thirteen votes to one with China abstaining. A Russian counter-draft failed in May 2024 on a seven to seven split. Two attempts, two failures, in the space of a month.

The Open-Ended Working Group on reducing space threats ran four sessions across 2022 and 2023 and could not agree a report. Its merged successor, the OEWG on the prevention of an arms race in outer space in all its aspects, spent its first session in April 2025 failing to agree an agenda and going partially closed. It managed an agenda by consensus in July 2025. It met again in Geneva in July 2026, with a mandate running to 2028. I could find no substantive outcome from that session.

And in November 2025, at the eightieth General Assembly, the United States and Israel were the only two states to vote against the annual PAROS resolution. It passed one hundred and seventy seven to two. The historical US position on that resolution was abstention.

That is the ground on which this announcement lands. Not a functioning regime with a new violation in it. A regime that has produced one binding sentence on weapons in space since 1967, and a diplomatic process that has not produced a consensus document in four years.

The strongest argument for the announcement, and why I do not buy it

Let me put the other case properly, because it deserves better than dismissal.

Deterrence requires the other side to know you can hurt them. Russia and China have been developing counterspace capabilities for two decades and have been considerably less forthcoming about it. The Secure World Foundation’s ninth Global Counterspace Capabilities assessment, published in April 2026, documents Russian rendezvous and proximity activity in the geostationary belt, a broad array of Russian ground-based jammers, and reporting that China may have a new direct-ascent interceptor, alongside repeated Chinese refuelling operations in GEO through the second half of 2025. Thomas González Roberts at Georgia Tech makes the fair point that the United States is at least taking a different approach from states that obfuscate.

On that reading, saying it out loud is honesty, and honesty reduces the chance of somebody misreading an ambiguous manoeuvre and starting something.

I do not think it survives contact with what was actually said. Clayton Swope at CSIS has the sharpest version of the objection and I think he is right: disclosure alone does not deter. Deterrence requires an adversary to understand both your capability and the conditions under which you would use it. This announcement supplied neither. Roberts himself notes that it tells you nothing about method of attack, reversibility, attributability, collateral damage or observability, which is to say nothing you would need in order to calibrate a response.

What it does supply, free of charge, is a talking point. Victoria Samson at the Secure World Foundation says the announcement crosses a line and risks normalising the weaponisation of orbit, among allies as much as adversaries. Cassandra Steer argues it validates exactly the accusation Moscow and Beijing have been making for years and strengthens their preference for a restrictive binding instrument over the Western responsible-behaviour agenda.

Watch what actually happened. Maria Zakharova’s response on 15 September did not only condemn the deployment. It attacked the moratorium on destructive anti-satellite missile testing as laden with hypocrisy. That moratorium, announced unilaterally by the United States in April 2022 and endorsed by a General Assembly resolution in December 2022 by one hundred and fifty five votes to nine, is the single most successful norm-building initiative the space security community has produced this century. Thirty eight states had joined it as of November 2024.

It is also the one thing standing between us and more events like Fengyun-1C. When China destroyed that satellite in January 2007 at around 850 kilometres, it created more than three thousand catalogued fragments, roughly two thirds of which are still up there. When Russia destroyed Cosmos 1408 in November 2021 at around 480 kilometres, it created about eighteen hundred, and almost all of them have decayed. Same act, different altitude, a difference of decades in consequence.

Handing Moscow a rhetorical stick to beat that moratorium with is not a small cost. It may turn out to be the most expensive part of this week.

What this means if you operate a satellite

This is the part that has been almost entirely absent from the coverage, and it is the part that matters to most readers of this journal.

The overwhelming majority of active satellites are commercial. The European Space Agency’s Space Environment Report, issued on 8 September 2026, models around 11,300 active payloads among 68,450 tracked objects larger than ten centimetres. A single commercial operator accounts for more than half of them.

Military dependence on that commercial layer is not marginal and it is deepening. SpaceX won a 2.29 billion dollar Space Force contract in May 2026 to build out the service’s low Earth orbit communications backbone, and holds a 4.16 billion dollar prime contract for the air moving target indication constellation whose first launches are due this month. Commercial space domain awareness providers are now part of the threat picture in their own right, with LeoLabs detecting Chinese counterspace manoeuvring in June 2026.

Here is the legal consequence, and operators should sit with it. Under the law of armed conflict, a civilian object that makes an effective contribution to military action by its nature, location, purpose or use becomes a military objective. The Woomera Manual on the International Law of Military Space Operations, published by Oxford University Press in 2024 under an editorial team led by Jack Beard, with Dale Stephens and David Koplow, states the position directly: third party civilian satellites that have become military objectives through their use for military purposes can be legitimate targets.

That is not a proposal. The Woomera Manual is descriptive of the law as it stands, drawn from state practice. It is worth being clear that this is a different exercise from the McGill Manual, often called MILAMOS, which is expressly confined to peacetime and does not address armed conflict at all. If you have seen the two cited interchangeably, they are not interchangeable.

The Woomera Manual also does something I find genuinely interesting. It holds that belligerents must select methods that minimise debris creation, because fragments endanger every user of the domain rather than only the adversary. That is a precaution in attack owed, in effect, to everybody. I am not aware of a close analogue anywhere else in the law of armed conflict, and if the manual’s position is right it is the single most useful legal development of the past decade for commercial operators.

Three practical questions follow.

Do you know what your capacity is being used for? A hosted payload, a leased transponder, a government customer buying imagery. The targeting test turns on use, and use can be established by your customer rather than by you.

Have you read your war risk exclusion recently? Expanding war risk exclusion language is one of the structural shifts reshaping a space insurance market that was already hard after the 2023 loss year. An in-orbit policy written against the geopolitical assumptions of 2019 is not priced for this. I cannot point you to public exclusion wording, because there is not any, which is itself the point. Ask your broker.

Who bears the loss if you are collateral damage? The Liability Convention gives you a fault standard for damage in orbit and never defines fault. There is no agreed rule of the road, no right of way, and no forum with compulsory jurisdiction. We have written about that gap before and it has not narrowed.

No commercial operator, industry association or insurer had said anything publicly about the announcement as of 17 September. I find that silence more interesting than any of the statements.

Where this leaves us

I do not think the United States has broken the Outer Space Treaty, and I think people claiming otherwise are reading paragraph two of Article IV as though it governed Earth orbit. It does not.

I also do not think that is reassuring, and I am uneasy with the number of lawyers this week who have treated the compliance question as the end of the analysis. A rule that permits everything the parties want to do is not a rule that is working well. It is a rule that has stopped being tested.

Article IV did one job, in 1967, superbly. It is still doing it. Everything that has happened in orbit since has been governed by nothing much at all, and the forum built to change that has spent forty years demonstrating that it cannot.

The question worth asking is not whether Washington complied. It is what the other one hundred and seventeen states parties intend to do about a treaty that turns out to permit this. So far the answer from Europe, which has the most commercial exposure and the least counterspace capability of any major bloc, has been nothing at all. No European government, no EU institution and no European Space Agency statement had been reported as of 17 September.

That silence is a choice, and it is the one I would watch.

Questions people actually ask

Did the United States break the Outer Space Treaty by putting weapons in orbit?

Almost certainly not. Article IV paragraph 1 prohibits placing nuclear weapons or other weapons of mass destruction in orbit around the Earth, installing them on celestial bodies, or stationing them in outer space in any other manner. It does not prohibit conventional weapons in Earth orbit. The United States asserts compliance and no official has identified the weapon type. The broader “exclusively for peaceful purposes” obligation in Article IV paragraph 2 applies only to the Moon and other celestial bodies.

What exactly did the US say and when?

Secretary of the Air Force Troy Meink said on 14 September 2026, in a keynote at the Air and Space Forces Association conference at National Harbor, Maryland, that the United States has on-orbit space control weapons capable of defending the Joint Force against hostile adversary action. He declined to give the weapon type, numbers, orbits, deployment date or testing history. A Space Force spokesperson confirmed and expanded on the statement the following day.

Does the Outer Space Treaty say space must be used for peaceful purposes?

Only in part, and the distinction is being widely misreported. The “exclusively for peaceful purposes” language in Article IV paragraph 2 applies to the Moon and other celestial bodies, together with bans on military bases, weapons testing and military manoeuvres there. Earth orbit is not covered by that paragraph. State practice since 1957 has read “peaceful” as meaning non-aggressive rather than non-military.

Can a commercial satellite be a lawful military target?

Under the law of armed conflict, yes, if by its nature, location, purpose or use it makes an effective contribution to military action and its destruction offers a definite military advantage. The Woomera Manual, published in 2024, states that third party civilian satellites that have become military objectives through use for military purposes can be legitimate targets, subject to proportionality and precautions in attack, including a precaution to minimise debris creation. The McGill Manual, or MILAMOS, is confined to peacetime and does not address this.

Is there any treaty that would prohibit conventional weapons in orbit?

No treaty in force does. The Russia and China draft Treaty on the Prevention of the Placement of Weapons in Outer Space has been tabled since 2008 and revised in 2014 but has never been negotiated, with the United States objecting that it fails to define a weapon in space, ignores ground-based threats and contains no verification mechanism. The UN Open-Ended Working Group on preventing an arms race in outer space has a mandate running to 2028.

What happens to debris if a weapon is actually used in orbit?

It depends overwhelmingly on altitude. China’s destruction of Fengyun-1C at roughly 850 km in January 2007 produced more than three thousand catalogued fragments, around two thirds of which remain in orbit. Russia’s destruction of Cosmos 1408 at roughly 480 km in November 2021 produced about eighteen hundred, nearly all of which have since decayed. Atmospheric drag clears low orbits within years and high orbits within centuries.