6 Tankers or 22? The Refueling Gap That Prices Every Mars Flight

A NASA safety panel says filling a Starship for the Moon could take around 22 flights. Elon Musk says 6 to 8. That is not a rounding error — it is two different futures for what deep-space flight costs.

The most important number in the Mars business this week was not a valuation or a launch date. It was a count of gas-station runs.

On Thursday, October 22, NASA's Aerospace Safety Advisory Panel held its quarterly public meeting, and panel member Charlie Precourt — a former NASA astronaut — flagged the fueling plan for SpaceX's Starship lunar lander as a major risk for Artemis IV, the crewed Moon landing planned for 2028. "We need to keep our eyes on numerous required fuelings for the lander," Precourt said. "I think the last estimate is around 22 flights."

Ars Technica's Eric Berger posted the remark on X. Within hours, Elon Musk replied: "That's way too high. Actual number will be more like 6 to 8."

Neither side has published a breakdown of its math. That silence is part of the story — because the gap between 22 and 7 is not an argument about arithmetic. It is an argument about how capable the next Starship actually is.

Why a Moon ship needs a fleet

Starship is so large that it burns nearly all of its propellant just reaching Earth orbit. It cannot simply keep going to the Moon, and it certainly cannot keep going to Mars. SpaceX's answer is a depot: a Starship configured as an orbital storage tank launches first, then tanker Starships fly one after another, dock with it, and transfer their leftover propellant. Only when the depot is full does the mission ship launch, dock, fill up, and depart.

Every tanker in that chain is a complete Starship launch — booster, ship, turnaround, weather, and risk included.

Three variables decide the count: how much propellant each tanker delivers to orbit, which rises sharply if the upgraded V3 performs as SpaceX expects; boil-off, because the cryogenic methane and oxygen slowly warm and vent while the depot waits; and transfer loss in an orbital fueling operation no one has yet demonstrated at this scale. Through those variables, the two numbers stop sounding like spin. The panel's 22 looks built from what Starship can lift today, with conservative losses. Musk's 6 to 8 looks built from the V3 still being proven. Only flight data settles it.

The count has also been drifting for years. In 2021, a government document suggested 16 launches; Musk called that "extremely unlikely" and put the maximum at eight tankers. In November 2023, NASA officials cited figures in the "high teens." Now the panel's estimate has moved up to around 22, while Musk's has not moved at all.

The timing sharpens it. The dispute landed the day before NASA's October 23 "Artemis acceleration" briefing. Artemis III, targeted for mid-2027, has been recast as an orbital docking test that needs no refueling flights. The campaign begins after: SpaceX is expected to fly its first tanker within a few months of Artemis III, then attempt an uncrewed lunar landing before the end of 2027. Berger's estimate shows the calendar cost: even at two launches per week by late 2027, filling a depot at the panel's number would take roughly three months. At 6 to 8 tankers, it fits inside a month.

There is a control group in the same program: Blue Origin's Blue Moon lander does not rely on orbital refueling for its initial mission design. If refueling proves slow or lossy, the architecture that avoids it gains ground. If it works at Musk's numbers, Starship's approach scales to payloads no single-launch lander can match.

The Money Angle

Follow the tanker count and you follow the money, because each tanker flight is a unit of cost, production, and risk multiplied by every deep-space mission.

At 22 flights per departure, a Moon landing is a small airline operation mounted for a single trip: more ships built or refurbished, more pad time, more dockings where a failure can strand a depot, and more months of boil-off eating cargo already paid to lift. Cadence becomes a hard requirement, not a bragging point. That is the world implied by SpaceX's request, now in FAA environmental review, to fly up to 456 Starship launches a year from Starbase — less ambition than arithmetic, if the panel is right.

At 6 to 8 flights, the economics invert: the depot fills in weeks, the same fleet supports more departures per year, and cost per landed mission falls toward something a program can repeat — the entire premise of Artemis as a sustained return.

For Mars, multiply again. Every Mars-bound Starship must also be refueled in Earth orbit, and departures compress into a transfer window that opens roughly every 26 months. A multi-ship fleet runs several campaigns in parallel, against a planetary deadline that does not move. At 22 tankers per ship, that demands a launch operation of a scale no one has ever run. At 7, a serious Mars campaign becomes an industrial problem instead of a fantasy. The Moon argument is the rehearsal — its winner quietly sets the price of the first Mars fleet.

What to watch is not the next exchange on X. It is the first orbital propellant-transfer demonstration, the first measured boil-off figures from a real depot, and the delivered payload of the V3 tanker. Those three data points will collapse a 14-flight argument into a fact — and with it, the first honest price tag for leaving Earth orbit at scale.

Gear tie-in

wltk SpaceX Starship 1/375 Diecast Rocket Model — a diecast model (Arts, Crafts & Sewing) of the full Starship stack, booster and ship together: the same two pieces that, in the real refueling architecture, have to fly again and again before one of them is allowed to leave.