🌿freegardner

Science

NASA's Lunar Landing Plan Now Rests on a Docking It Won't Let Astronauts Enter

21 Nov 2026 · via Feeds.arstechnica

NASA's Lunar Landing Plan Now Rests on a Docking It Won't Let Astronauts Enter
Image: NASA (Public Domain)

A Catch That Changes the Arithmetic of Going Back

A rocket booster caught in mid-air by the arms of its own launch tower sounds like the kind of stunt that belongs in a highlight reel, not a mission plan.

Last weekend SpaceX caught its Starship rocket at the launch tower, the first recovery of its kind. For the company’s own purposes, the recovery teaches engineers how to reuse Starships quickly, which in turn speeds the deployment of its newest generation of Starlink Internet satellites. For NASA, the payoff is subtler and more consequential. A higher launch cadence means more flights, and more flights mean faster learning as SpaceX develops a deep space version of Starship meant to carry crews to the surface of the Moon.

The infrastructure is being laid in parallel. Early next year, SpaceX plans to bring a second Starship launch site online at Kennedy Space Center in Florida. A second Starship factory is under construction at Kennedy as well, joining the existing production line at Starbase, Texas.

The model that now needs revising is the old assumption that a rocket company proves a vehicle once and then flies it. Here the vehicle is proven in increments, and each increment feeds the next. Saturday’s catch was not the destination.

The Refueling Test That Everything Else Waits On

SpaceX’s propellant transfer demonstration has slipped two years since late 2024 — effectively a day-for-day delay, because the company was ironing out technical problems with the vehicle itself. The plan is now for the first quarter of next year: two Starships, one launched from Florida and one from Texas, docking together in low-Earth orbit and moving a “meaningful” amount of propellant from one ship to the other. Steve Creech, the program manager for NASA’s Human Landing System, calls that transfer a big learning point, because the entire architecture depends on propellant aggregation and multiple propellant flights. [1] The test would mark the first time one spacecraft has offloaded a significant quantity of cryogenic propellants — in this case, methane and liquid oxygen — onto another vehicle in space. It is the culmination of decades of government and commercial research on cryogenic orbital refueling. The results will tell engineers how much propellant boils off during the transfer. That single unknown is why the number of tanker flights needed for a lunar mission remains so fuzzy.

The estimates span a wide canyon. The low end is six to eight refueling flights to support a Starship lunar mission, a figure from SpaceX founder and CEO Elon Musk that depends on future Starship upgrades reaching the vehicle’s full potential. NASA’s own conservative planning figure is 20 refueling launches. For context, a Starship carries 1,600 metric tons of propellant with its tanks full, and refueling flights using future fully-optimized Starships might fill those tanks with as few as six tankers — which is not the case today. SpaceX’s goal is to lift 100 to 200 metric tons of payload to low-Earth orbit. Today’s version, Starship V3, is still far short of 100 metric tons.

A taller, heavier Starship V4 is in work, flying with nine Raptor engines instead of the six-engine variant flying now. Musk said in 2025 that V4’s first launch was targeted for 2027, and there has been no recent update. The initial Artemis lunar missions will use Starship V3. The refueling test is therefore a rehearsal performed with the equipment that will actually fly, not a placeholder.

NASA's Lunar Landing Plan Now Rests on a Docking It Won't Let Astronauts Enter (Image 1)

A Crew

Cabin and a Docking Port NASA changed the Artemis III flight plan earlier this year, rerouting it from a lunar mission to a low-Earth orbit mission. A four-person crew launches on a Space Launch System rocket inside an Orion spacecraft. Orion docks first with Blue Origin’s Blue Moon lander for two days of tests, after Blue Moon launches on a New Glenn rocket. The astronauts open hatches and enter Blue Moon’s pressurized crew cabin to test its environmental control and life support systems — the systems that keep a crew alive in space. Then they return to Orion, undock, and set up a rendezvous and docking with SpaceX’s Starship in a slightly lower orbit, kept below most of the Starlink constellation. Orion stays attached to Starship for just two hours. The astronauts never enter the SpaceX vehicle.

The asymmetry is deliberate. Blue Origin no longer requires in-space refueling for Artemis landing missions, and it has comparatively little experience with life support systems; its suborbital space tourism flights were brief and demanded far less than longer durations require. Artemis III will be Blue Origin’s first time flying a crew cabin for orbital spaceflight. SpaceX, by contrast, has extensive life support experience from its Dragon crew spacecraft. The bigger challenges for SpaceX were the prop transfer and reaching the uncrewed landing demo.

Creech frames the split plainly. With Starship, NASA learns about integrated stack control, because when Orion docks with Starship, Starship will have control of the attitude and control of the stack — a large vehicle doing the steering. With Blue, Orion will control the stack for this mission, a risk-mitigation choice; a later mission will require the fully loaded Blue lander to control it. “If we added a crew cabin, that would slow us down on the other things and not add to what we’re already doing,” Creech said of SpaceX’s part. [1] The Starship on Artemis III flies with a full complement of Raptor engines — the engines that power the vehicle — and a docking port on its nose. Blue Moon arrives with no main engine, only smaller thrusters for pointing control and minor maneuvers, though it carries the crew cabin.

NASA added milestones to both contracts. SpaceX must build a prototype crew compartment called Cabin One at its Starbase, Texas test site to check parts of its crew cabin on the ground. Blue Origin must hotfire a three-engine Blue Moon propulsion unit inside a vacuum chamber at NASA’s In-Space Propulsion Facility in Ohio. Blue Origin plans its first Blue Moon test mission, Mark 1, early next year after a monthslong delay caused by a rocket explosion on its launch pad in May; that version has a single hydrogen-fueled BE-7 main engine, while future human-rated landers will have three.

The architecture beneath all of this has been redrawn. The previous plan required a 90-day loiter in lunar orbit for Orion to arrive, a requirement that drove complexity and made the lander job hard for both companies. That loiter is gone. SpaceX solved its version by proposing an Earth orbit rendezvous with Orion and taking Orion to the Moon. Both companies now have a four-day initial surface stay requirement, down from six and a half days. The old figure was baked in because Orion was to settle into a near rectilinear halo orbit — a long, looping path around the Moon — predicated by the expected location of the now-canceled Gateway lunar space station, which in turn followed from Orion’s inability to reach a closer orbit. A new stretched Centaur upper stage for SLS should give the rocket the energy to send Orion to low-lunar orbit. NASA has not determined whether that stage will be ready for Artemis IV; if not, Artemis IV could fly with the last of the in-stock SLS upper stages. NASA is also still deciding which lander Artemis IV will use.

What remains unresolved is where exactly Starship meets Orion. A graphic presented by Jeremy Parsons, NASA’s Artemis program manager, illustrated a “final tanking orbit” before Starship docks with Orion in high-Earth orbit. Creech is direct: “We’re still working that out. We’re making progress, but where we actually rendezvous with Orion is still in trade, so that’s why you saw a little bit of fuzziness there.” [1] The question underneath is whether SpaceX must refuel Starship at an altitude somewhere above low-Earth orbit. That is the one variable the whole plan hangs from, and it is still being traded.

Robots First, or Pilots First

Every lander must complete an uncrewed landing demonstration before a crew flies. Those demos are targeted for late 2027 or early 2028, but they will not test everything. Starship’s elevator — the lift that would carry astronauts from the cabin down to the lunar surface — is not part of the uncrewed landing demonstration; engineers will check its performance in a simulated lunar environment at a test lab at Marshall Space Flight Center in Alabama. NASA did not initially require either company to prove its lander could take off from the Moon before a crew mission. Officials later added the lunar takeoff requirement to both uncrewed demos.

The two companies’ takeoff plans differ. Blue will land and then do a full ascent back to the rendezvous point with Orion. SpaceX will land, take off again, get to main stage on the engines, and then land a second time on the surface. That gives SpaceX two shots at proving Starship can land safely on the Moon, at a cost. “When I came into the program, I was like, ‘Why are we not going all the way to the rendezvous point?’” said Creech, who took over as HLS program manager in 2025. “Some of my key people felt strongly we would get more out of the way we’re doing it now. There are others that have different opinions… We tend to have a lot of technical discussion and debate. That point will continue to be debated.”

NASA's Lunar Landing Plan Now Rests on a Docking It Won't Let Astronauts Enter (Image 2)

A parallel effort is running on the same question from a different house. China’s program plans to land people on the Moon by 2030, and its published sequence mirrors NASA’s on paper while skipping crewed lunar test flights entirely. Chinese space officials laid out the approach at the 7th Manned Space Symposium in Beijing on Oct. 31. A crew flies the Mengzhou, or “Dream Vessel,” capsule to Earth orbit in 2027. Then the robots take over: the Lanyue, or “Embracing the Moon,” lander touches down on its own in 2028, and an uncrewed mission rehearses the full landing sequence in 2029. “For any crewed mission, we will try to fly it in an unmanned state first,” said Wang Xiang, chief designer of the Chinese lunar exploration program.

The contrast is institutional, not just technical. China’s military-run program trusts machines to prove hardware; NASA’s civilian program, working in public view, leans on human test pilots. NASA Administrator Jared Isaacman described the American approach at an Oct. 23 news conference as the Apollo playbook: crews on key test flights, frequent launches, each mission building on the last. “It’s not entrepreneurial thinking or tech industry pixie dust or some new management philosophy,” he said. “It’s a playbook this agency knows really well — the same playbook that went from Apollo 7 splashdown to Apollo 11 in just nine months.” During Apollo, NASA’s lunar lander never landed on the Moon before Neil Armstrong and Buzz Aldrin descended to the Sea of Tranquility; the agency did fly a dress rehearsal around the Moon on Apollo 10 that included a test of the lunar module’s landing abort capability.

NASA’s Human Landing System office now holds two contracts, two philosophies of rehearsal, and one unresolved variable — the orbit where the tanker meets the ship. The first crewed lunar landing of the Artemis program is now targeted for Artemis IV, scheduled to launch before the end of 2028. Before that happens, the debate Creech describes will have to resolve into a number.


Sources

  1. Ars Technica — Quote source (original article)

Mentioned organisations (context, not sources)

← back to the garden