A NASA concept reported in late October 2026 lays out two orbital propellant transfers for SpaceX’s Starship Human Landing System (HLS): first at a depot in low Earth orbit (LEO), then at a final tanking orbit (FTO), before Starship heads to meet Orion in high Earth orbit.

How the Artemis IV refueling sequence would work

Flow diagram of a proposed Artemis IV sequence: a depot in low Earth orbit supplies Starship HLS twice before its rendezvous with Orion.

Multiple Starship tankers would carry liquid methane and liquid oxygen to a depot in LEO. Uncrewed HLS would dock with the depot and receive propellant. It would then raise its orbit to the FTO. After the depot was replenished, it would meet HLS for a second transfer. Starship would then head to high Earth orbit to rendezvous with Orion.

Where refueling fits in the crewed mission

The reported sequence places the Space Launch System (SLS) launch of Orion before HLS’s final transfer, a staging arrangement that could provide more flexibility in SLS launch timing. The mission outline sends Orion directly to low lunar orbit rather than the higher orbit used in earlier plans, reducing Starship’s fuel needs.

Four astronauts would travel aboard Orion. Two would remain there while the other two descend to the lunar surface in Starship. For the broader mission timeline, see NeoTeo’s earlier Artemis IV coverage.

What the tanker-flight estimates say

The concept calls for multiple tanker flights but does not specify a count. A Government Accountability Office assessment from summer 2021 was reported as estimating that the architecture could require up to 16 launches. In October 2026, Charlie Precourt, a member of NASA’s Aerospace Safety Advisory Panel, was reported to estimate as many as 22 flights; Elon Musk’s estimate was reported as six to eight.

A Reddit post reproducing NASA Artemis Program manager Jeremy Parsons’s remarks from an October 23, 2026, Artemis Acceleration Q&A attributes to him the view that NASA was not tracking as high as 22 under the assumed transfer performance. Parsons said six to eight was probably closer to SpaceX’s planning estimate and that an orbital propellant-transfer demonstration would be the best way to determine the required number of flights.