NASA has committed $20 billion to building a permanent lunar base. The first crewed landings are targeted for 2028, on a schedule of two missions per year. The hardware exists, or is being built. What does not yet exist, in any documented form, is the operational software layer that would run it.
That is the argument Laura Crabtree makes in a SpaceNews op-ed published this week. Crabtree is the CEO of Epsilon3, a company that sells mission execution software to space and defense organizations. She spent a decade at SpaceX working on Dragon mission operations. Her pitch is self-interested, but the structural observation behind it is not wrong.
"The hardware plan is funded," Crabtree writes. "The software strategy is not." The Artemis architecture, as laid out by NASA Moon Base Program Manager Carlos Garcia-Galan at the Ignition conference in late February, calls for 25 launches and 21 landings through 2028, establishing what Garcia-Galan called "the beginning of operations" at the lunar south pole. That is not a test campaign. It is a production schedule.
ISS took years to develop its operational rhythm. The lunar surface program is attempting to establish one in roughly two years, with more providers, more launch sites, and more simultaneous mission threads than any ISS era program managed.
The problem Crabtree identifies is not hypothetical. Space operations at scale require software that coordinates manifests, tracks subsystem health across multiple simultaneous spacecraft, manages downlink data pipelines, schedules crew activities, and integrates with ground networks across variable-delay communications environments. These are not problems a launch vehicle solves. They are problems a mission operations infrastructure solves.
NASA's recently announced workforce directive is part of the picture. The agency has directed civil servant staff to rebuild in-house development competencies and work side-by-side with Artemis partners. That is a people and process decision. The software architecture decision, Crabtree argues, has not been made in public.
The conflict of interest is worth stating plainly: Crabtree sells mission execution software to space organizations. Her op-ed describes a problem her product category addresses. That does not make the problem unreal. It does mean the proposed solution should be held to a high bar before anyone treats it as the answer.
What is known is that the hardware plan has passed the point of political review. The Lunar Gateway has been cancelled. The $20 billion Ignition commitment through 2032 is real. Semiannual landings starting 2028 are the target. Garcia-Galan has said Phase One is "not a test, it is the beginning of operations." An operational program that size, with multiple providers, needs coordination software that does not yet have a public architecture.
The gap is real even if the solution Crabtree proposes is not the only one. Whether NASA procures Epsilon3, builds something in-house, or adopts a legacy system from the ISS program, the missing ops layer is the next infrastructure question the agency needs to answer now that the hardware question has been answered.
