Clocks on the Moon run ahead. Not because they are poorly built, but because the gravitational field there is weaker, and time passes faster under weaker gravity. An atomic clock on the lunar surface gains roughly fifty-six millionths of a second per day against an identical clock on Earth, varying with location and altitude.

That sounds like nothing. It is the time light needs to travel seventeen kilometres. For anyone navigating, landing or docking, those millionths decide between an arrival and an impact. Without a shared basis of time, position errors grow every day, and they grow a little differently for every vehicle, every transmitter and every ground station.

So the Moon is getting its own time. That much is settled. What is not settled, and what actually interests me here, is a different question entirely: who gets to define it.

The quietest successful institution we have

Time is not a constant of nature that we read off. Time is an agreement about what simultaneous means, and an agreement needs someone to hold it.

On Earth, a treaty holds it. On 20 May 1875, seventeen states signed the Metre Convention in Paris, creating the International Bureau of Weights and Measures. Coordinated Universal Time is still computed under the authority of the General Conference on Weights and Measures, from the data of hundreds of atomic clocks in dozens of countries. It is controlled by no single country, no ministry and no company. No corporation can buy it, no president can change it by directive.

I consider this the quietest successful institution we have. It governs a quantity on which every exchange, every power grid, every navigation system and every data link depends, and almost nobody knows its name. It has worked for a hundred and fifty years precisely because no one owns it.

That order is now being claimed from two directions at once. The first lies on Earth. Since 1972, world time has been aligned with the planet's actual rotation by occasionally inserting leap seconds. In 2022, member states agreed to loosen that coupling by 2035 at the latest and to allow a larger divergence between astronomical and atomic time. Meanwhile Earth has spun up enough that a negative leap second may become necessary for the first time, something for which no tested practice and no software experience exists anywhere. Specialists put the probability before 2035 at thirty to forty per cent. In October 2026, replacing the leap second with a far rarer correction comes to a vote.

We are, in other words, detaching time from the rotation of our planet. It is turning from an astronomical fact into a pure convention. And this is happening at the very moment we first have to extend it to a second celestial body.

Three definitions, three holders

Three answers to lunar time are currently running in parallel, and they differ not in their physics but in who stands behind them.

The first is national. On 2 April 2024, the White House Office of Science and Technology Policy directed NASA to develop Coordinated Lunar Time and to deliver an implementation strategy by the end of 2026. Four requirements are fixed: traceability to Coordinated Universal Time, accuracy sufficient for precision navigation, resilience in case of lost contact with Earth, and scalability beyond the Earth and Moon system. This is an executive instruction from one state for a body that belongs to no state.

The second is already finished, and it comes from the other bloc. Researchers at the Purple Mountain Observatory of the Chinese Academy of Sciences have published a ready-to-use software package that computes a lunar coordinate time and converts it into the barycentric time scales, with an accuracy better than a sixth of a nanosecond out to 2050. It is openly available. Anyone planning a lunar mission today can use it without waiting for any conference.

The third is multilateral and slow. The consultative committee responsible for time and frequency is holding workshops with lunar system providers and preparing recommendations. Among the draft resolutions for the 2026 General Conference is a text on defining an international lunar reference time scale and its traceability to Coordinated Universal Time. In parallel, working groups at the International Astronomical Union, the International Association of Geodesy, the satellite navigation committee and the telecommunication union are addressing the same question.

Three definitions, three holders: a government, a national observatory, a treaty body. Physically they will converge, because relativity respects no borders. Institutionally something else is being decided, namely whether the first infrastructure beyond Earth answers to a common holder or to whoever ships first. One of the physicists involved named the risk without decoration: without agreement, we face a time zone war in space.

Where the model breaks

At the Moon you can still pretend to share a present. A radio path takes a little over a second, and a conversation is sluggish but possible. Everything conceived for the Moon can therefore still be thought of as an extension of terrestrial order: a centre defines, a relay distributes, a receiver follows.

At Mars that model breaks in three places at once, which is why Mars is the real test case.

The first break is the magnitude. Physicists at the American standards institute have calculated that clocks on Mars run on average four hundred and seventy-seven millionths of a second per day faster than on Earth. What matters is not the number but its restlessness. Across a Martian year that lead varies by two hundred and twenty-six millionths of a second per day, so in practice between three hundred and sixty-four and five hundred and ninety, because the Martian orbit is markedly more eccentric and its distance from the Sun varies more strongly. On top of that sits a slower swing of some forty millionths, building over seven synodic cycles. The lunar offset is nearly a constant and can be fixed once. The Martian offset breathes. A breathing quantity cannot be decreed; it has to be modelled continuously, and someone has to be answerable for which version of the model is currently in force. That calls for an operation, not a decision.

The second break is simultaneity. Depending on orbital geometry, three to twenty-two light minutes separate Earth and Mars. That is not a delayed shared present but the end of one. Any statement about what is happening on Mars right now is, from here, a convention rather than an observation. The planet also keeps its own beat: a Martian day lasts twenty-four hours, thirty-nine minutes and thirty-five seconds. Anyone living by it drifts a little further from every earthly calendar each day.

The third break is scheduled. Roughly every two years, Earth and Mars stand on opposite sides of the Sun, whose hot plasma renders radio signals unusable. NASA halts the sending of commands entirely for about two weeks. The vehicles work through a list loaded in advance and report only their state of health. Mars is therefore regularly, predictably and unavoidably ungovernable from Earth. Anything that would need an Earth decision inside that window is simply undefined.

What has to be built

Those three breaks produce a build list, and it is more concrete than the current debate suggests.

Mars needs its own coordinate time and a surface scale defined against its own reference level, not against Earth's. The formal framework has existed in the astronomical literature for years. What is missing is the implementable version: a continuous derivation from reference systems through operational clock budgets, light-time observables and validation procedures, the work that turns an equation into an instrument. That necessarily includes a published rate model rather than a constant, because the lead changes with the planet's season.

Mars needs a local clock ensemble, on the surface and in areostationary orbit, so that its time is generated locally rather than derived from Earth. The fourth requirement written for the Moon, resilience in case of lost contact with Earth, is a precaution there. At Mars it is normal operation.

Mars needs an interoperable relay and navigation layer. For the Moon this is already under way: the European space agency is building its lunar programme to jointly agreed standards, with initial operations at the end of 2028 and full service in 2030. For Mars there is no equivalent. Data return depends on an ageing fleet belonging to individual agencies, and each failure tears a hole that nobody fills.

And Mars needs a civil reckoning of time that people can live by: the relation of sol to second, a local time per site, a calendar. On Earth that question took an international treaty and half a century of argument. For Mars, nobody yet holds the mandate even to pose it.

The confirmation horizon

Behind all of this lies a pattern that reaches well beyond timekeeping, and I want to give it a name.

There is a distance beyond which asking back stops being an option, because the answer arrives later than the moment that needed it. I call this the confirmation horizon. On this side of it, order can be defined centrally and adjusted as you go, because you can always ask. Beyond it, only what was decided in advance has any effect.

The Moon sits just inside. Mars sits clearly outside, not in an emergency but in daily operation. And because time is the precondition of every other form of coordination, it is the first case where this bites. It will not be the last. Every rule about safety, about jurisdiction, about how to treat what is found out there faces the same condition.

That has an uncomfortable consequence for institutions. A body that acts only when asked is powerless beyond the confirmation horizon, however well staffed it may be. Authority out there must have been transferred beforehand, or it does not exist. This raises the demands on what we settle now considerably, because the later correction that earthly politics habitually relies on is not available.

What is actually on the table

Back to October 2026. It looks like a piece of metrological housekeeping, and it is being reported that way. I read it differently.

The fourth requirement of the American directive explicitly demands that the lunar standard be scalable beyond the Earth and Moon system. That says openly what is happening: whoever writes lunar time writes the template for Mars and for everything after it. This is not about the Moon.

The objection to the treaty model is obvious and deserves to be taken seriously. National programmes are fast, treaty bodies are slow, and a standard nobody implements is worthless. Whoever operates working clocks first sets the pace, whatever is resolved later. That is precisely why the race began while the deliberations are still running.

I still think the opposition is drawn in the wrong place. Implementation will be national anyway, carried out by agencies and companies with their own clocks and their own relays. Only the definition has to be common: the question of which scale everyone computes against, and who maintains it. For that we already have the one institution that has ever proven capable of the job, and it has done it for a hundred and fifty years without any state managing to capture it. Bypassing it because it is inconveniently slow would mean anchoring the first multiplanetary infrastructure nationally, in the one place where nationality has ceased to mean anything.

We will not be able to take this decision back. Whatever is set as convention now will be adopted by every mission that follows, because nobody switches the basis of time mid-operation. It is the first decision of order that applies to two celestial bodies at once, and it falls in a room otherwise reserved for kilograms and kelvin.

Which rule in your organisation would still hold if the query takes forty minutes to arrive and the answer stays away for two weeks?

Website: https://planet-futures.org