NASA's Safety Panel Warns About Artemis Refueling: About 22 Starship Launches, or Six to Eight? Musk and the Panel Disagree.

NASA's Aerospace Safety Advisory Panel (ASAP) warned at its Oct. 22, 2026 public meeting of "acute technical risks" in the plans to land astronauts on the Moon, pointing in particular to unproven in-space transfer of cryogenic propellant. A panel member cited a number of roughly 22 Starship tanker launches. SpaceX CEO Elon Musk said that was far too high and that six to eight is closer. This article lays out what each side said and why the number matters.

What the Panel Said

According to SatNews, which covered the meeting, ASAP warned of acute technical risks in the Artemis lunar landing plans and singled out the "unproven in-space cryogenic propellant transfer" they depend on. Panel members said dozens of rapid-turnaround cryogenic transfers would add compounding failure modes to crewed lunar flights.

Panel member Charlie Precourt, a former astronaut, said current NASA planning estimates suggest SpaceX may need about 22 Starship tanker launches to fully fuel an orbital propellant depot before a lander can leave for the Moon, according to SatNews. The Register, covering the same meeting on Oct. 23, describes a reported suggestion of about 22 launches linked to SpacePolicyOnline, and says it did not find an explanation of how the figure was calculated. We could not find one either.

The panel plans a formal recommendation that NASA re-examine mission objectives and system architectures for Artemis IV to reach a more balanced risk posture, SatNews reports. The Register says a written report is due early next year.

What Musk Said

Musk called the figure "way too high" on X and said the "actual number will be more like 6 to 8," according to The Register. The coverage we read gives no detailed reasoning from him, and we did not find a separate NASA response to the 22-launch estimate.

Why Refueling Is the Hard Part

Landers for Artemis need to be filled with propellant in orbit, so extra launches are needed just to deliver fuel. The unresolved problem is demonstrating reliable cryogenic transfers between spacecraft at the required scale. The Register notes that Apollo's lunar module needed only a single Saturn V launch because it used propellants that do not need to be kept extremely cold. Cryogenic propellants boil off in space, which means timing matters: the more launches needed to fill a depot, the more of the fuel is exposed to boil-off and the more chances for something to go wrong.

Where the Schedule Stands

Per SatNews, Artemis III was revised to add an Earth-orbit test mission in mid-2027. It would test automated rendezvous with SpaceX's Starship lander and Blue Origin's Blue Moon lander in low Earth orbit before astronauts head to the lunar south pole. The first crewed lunar landing has moved to Artemis IV, which SatNews dates to early 2028. Other coverage in our search results dated that landing to late 2028, so the target date is not settled in the sources we read. NASA Administrator Jared Isaacman is preparing an agency-wide Artemis update, SatNews says, and program managers must resolve the refueling bottleneck to hold the 2028 target.

ASAP also raised concerns outside the Moon plan, according to The Register: SpaceX's plans to phase out Falcon and Dragon could complicate extending International Space Station operations and launching its deorbit vehicle, and Boeing's Starliner earliest possible uncrewed launch is now mid-January 2027.

How to Read the Disagreement

This section is analysis, not reporting.

The two numbers differ by about a factor of three, and they matter for different reasons. The count of launches sets cost and schedule: each launch is a flight to prepare, fly, and recover. It also sets the risk, because every additional transfer is another chance for a failure. If the true number is closer to six or eight, the plan is much more plausible. If it is near 22, the cadence required is a major test of whether a reusable launch system can fly that often.

Neither side has shown its math in the coverage we read. Musk is the person who builds the vehicles, and the panel is an independent safety body, so each has a different incentive and a different kind of authority. The way to resolve it is a flight test of propellant transfer, which has not been completed at scale according to the panel's warning. Until then, the number is an estimate, not a measurement.

What We Cannot Conclude

We cannot confirm that 22 is NASA's official planning figure, how it was derived, or what NASA plans to do in response. We cannot verify Musk's number. The Artemis IV date differs between sources. We did not attend the meeting or read ASAP's materials.

Why It Matters

• The Moon plan depends on a technique not yet proven at scale. In-space cryogenic transfer is the central technical risk.
• A threefold gap in the launch count is a threefold gap in cost and risk. The disagreement is public and unresolved.
• The timetable is political as well as technical. The Register notes a desire to land astronauts before the current president leaves office.
• Watch for NASA's Artemis update and ASAP's written report. Both are expected, and either could change the numbers.