February 25, 2026. A 322-foot rocket crawled back to its hangar at 1 mile per hour, hauled by a crawler burning 165 gallons of diesel per mile. It had been on the launchpad for over a month. It had survived a liquid hydrogen leak, a wet dress rehearsal, and a second wet dress rehearsal. Then engineers found an interrupted helium flow to the upper stage. And just like that, the March window was gone.
Artemis II will launch no earlier than April 1, 2026. Due to trajectory requirements, the mission's launch windows open for only a few days roughly every four weeks. The helium issue means the mission missed the March 6-11 window. The next opportunity runs April 1-6. Miss that, and the crew goes back to Houston again. Reid Wiseman, Victor Glover, Christina Koch, and Jeremy Hansen have already been through quarantine twice. They are, by any reasonable measure, the most patient people on the planet right now.
I want to be clear about what actually failed here, because the coverage often muddles it. The rocket didn't blow up. Nobody got hurt. In addition to the initial hydrogen leak, a valve associated with Orion's crew module hatch pressurization required retorquing. A second wet dress rehearsal on February 19 was successful. Then, on February 21, a helium flow issue was observed, triggering the rollback. These are pre-flight test systems doing exactly what they are designed to do: catch problems before humans are sitting on top of the rocket. The system worked. The bureaucracy around it, however, is a different story.
The Same Problem, Twice
Here's what makes this maddening. Artemis I encountered the same issues ahead of its launch. "When you are experiencing some of the same issues between launches, you probably got to take a close look at your process for remediation," said NASA Administrator Jared Isaacman. He's right, and credit to him for saying it out loud. When the same helium and hydrogen headaches show up across two missions separated by over three years, that's not bad luck. That's a systemic process failure. It means the lessons weren't institutionalized. It means the turnaround between flights was so long that institutional memory degraded.
Three years between flights. Think about that. SpaceX launched Falcon 9 more than 90 times in 2023 alone. Iteration speed is how you fix problems. Infrequent launches are how you forget you had them.
The cost numbers make this worse to look at. NASA's spending on its Artemis program will reach a total of $93 billion by 2025, according to an audit by the NASA Office of Inspector General. The OIG projects the current production cost of a single SLS/Orion system at $4.1 billion per launch. For context: a SpaceX Falcon 9 launch averages about $67 million, translating to roughly $1,400 per kilogram to low Earth orbit. NASA's SLS has launch costs estimated around $2 billion per flight, pushing cost per kilogram dramatically higher. The SLS is, without exaggeration, the most expensive rocket flying today. One factor in the high cost of Artemis is that other than the Orion capsule, all components are expendable and single-use, unlike emerging commercial spaceflight systems. You build it, you fly it once, you throw it away. Every time.
Vera Santos would say fix Earth before burning $4 billion to go look at the Moon. I understand the argument. I just think it misses what's at stake in the next decade of spaceflight architecture.
Now Look at What NASA Just Did
While that 322-foot rocket was crawling back to the VAB, Isaacman held a press conference and announced something that actually matters more than this particular delay. NASA announced a major overhaul to its Artemis moon program, a "course correction" that will add missions and increase the pace of launches ahead of a targeted lunar landing attempt in 2028. The architecture is changing. NASA will attempt to launch Artemis III by mid-2027 to conduct key technology demonstrations in low-Earth orbit, including rendezvous and docking tests with commercially built lunar landers. After that, Artemis IV will launch in 2028 to land on the moon.
This is not a retreat. This is what good engineering looks like. Apollo didn't skip steps either. You test, you verify, you go. "The entire sequence of Artemis flights needs to represent a step-by-step build-up of capability, with each step bringing us closer to our ability to perform the landing missions. Each step needs to be big enough to make progress, but not so big that we take unnecessary risk given previous learnings," said NASA Associate Administrator Amit Kshatriya. That's not a bureaucrat covering himself. That's an engineer speaking plainly.
And notice what's embedded in this new architecture: SpaceX and Blue Origin are being asked to accelerate their lunar landers. Artemis IV and V would take astronauts to the surface of the moon, utilizing lunar landers designed and developed by SpaceX and Blue Origin. The commercial engine is being engaged more deeply, not less. That's the right direction. Private iteration speed, applied to the hardest part of the problem.
The Crew Makes This Worth It
None of the cost arguments change what Artemis II actually is. The 10-day mission will carry NASA astronauts Reid Wiseman, Victor Glover, and Christina Koch, along with Canadian Space Agency astronaut Jeremy Hansen, on a free-return trajectory around the Moon and back to Earth. Glover will become the first person of color, Koch the first woman, and Hansen the first person not from the United States to leave Earth orbit and to travel around the Moon.
The flight is set to take the crew farther from Earth than any previous crewed mission, before reentering Earth's atmosphere at a record speed of approximately 25,000 miles per hour. Twenty-five thousand miles per hour. These four people will travel farther from Earth than any human beings since Apollo 17 in 1972. That's 54 years of waiting. A helium valve is not going to stop that from being extraordinary.
The SLS is expensive. The delays are real. The model that built this rocket needs to be replaced by something that iterates faster and costs less. All of that is true. And on the day that rocket finally lights and those four people clear the tower, every one of those criticisms will still be true, and none of them will feel like they matter.
We are going. The only question is when.
