Report: NASA’s Management of the Human Landing System Contracts

Key Takeaways

• NASA has held HLS contract growth low, but schedule pressure has become the bigger problem.
• SpaceX and Blue Origin face hard technical barriers in fueling, integration, and flight readiness.
• The audit argues crew survival planning still lags behind the ambition of the Artemis campaign.

A Close Reading of NASA’s Human Landing System Audit

NASA has spent years presenting Artemis as the program that will return astronauts to the Moon and then keep them there often enough to build lasting operational experience. The NASA Office of Inspector General audit examines the part of that campaign with the least room for failure: the Human Landing System , or HLS. That is the vehicle class that must carry astronauts from lunar orbit to the surface, support them during their stay, and return them safely to orbit. The audit is not a broad reflection on Artemis politics. It is a focused assessment of contracts, management structure, technical progress, and crew safety.

The document centers on two providers. SpaceX is building a lunar version of Starship for the earlier Artemis surface missions. Blue Origin is developing Blue Moon for a later sustaining mission. NASA chose a service-based model rather than the older pattern of owning the spacecraft outright. That choice has shaped nearly every finding in the audit. It has constrained cost growth better than many large aerospace efforts. It has also left NASA dealing with the consequences of fast-moving private development cycles, evolving interfaces, and a safety case that is still being assembled while the vehicles themselves remain unfinished.

final-report-ig-26-004-nasas-management-of-the-human-landing-system-contracts.pdfDownload

What the Audit Set Out to Examine

The Inspector General organized the audit around three practical questions. Are the providers meeting cost, schedule, and performance goals. Is NASA’s insight and oversight model working. Is the agency identifying and reducing astronaut safety risks at a level that matches the danger of a lunar landing mission.

That framing matters because it keeps the report from drifting into broad institutional criticism. On cost, the picture is better than many critics of Artemis might expect. On schedule, it is much worse. On safety, the report is not saying NASA has ignored risk. It is saying NASA has not yet matched the depth of its safety preparation to the novelty and fragility of what it plans to fly.

The Contract Model Behind the Program

NASA used firm-fixed-price, indefinite-delivery, indefinite-quantity contracts with milestone payments. In plain terms, that means NASA pays for completed outcomes rather than reimbursing contractors for their labor as work unfolds. The logic is familiar from the Commercial Crew Program . Put more cost responsibility on industry. Let providers own their systems. Preserve competition. Reduce the government’s administrative burden. Reward delivery rather than effort.

The HLS path began under NextSTEP-2 . In 2020, NASA funded early lander work by Blue Origin, Dynetics, and SpaceX. In 2021, NASA selected SpaceX for the first full demonstration contract tied to Artemis III. In 2022, NASA extended SpaceX’s role for Artemis IV. In 2023, NASA awarded Blue Origin a separate contract for Artemis V. By the time of the audit, SpaceX’s potential contract value stood at about $4.3 billion and Blue Origin’s at about $3.1 billion.

There is a real policy argument underneath those numbers. Some observers prefer government-owned systems because they appear to offer tighter control. The audit supports the opposite conclusion on one narrow but important point. This commercial-style structure has been better at containing contract growth than a traditional cost-plus approach likely would have been. That is not a small result. NASA has struggled for decades with flagship programs whose costs rose with every schedule slip. Here, the government has absorbed delays without seeing proportionate contract expansion.

Cost Control Worked Better Than Expected

The strongest positive finding in the audit is financial. As of December 2025, the combined growth in the SpaceX and Blue Origin HLS contracts was only $266 million. SpaceX’s potential value rose by about 6 percent, or roughly $253 million. Blue Origin’s increased by less than 1 percent, about $13 million. For an effort of this size and difficulty, those are restrained numbers.

NASA also used schedule delay negotiations to obtain more value. When SpaceX requested a delay, NASA secured a lunar ascent test during the uncrewed demonstration mission. When Blue Origin needed more time, NASA gained additional testing, reviews, and data. That is a sharper outcome than the usual pattern in major aerospace contracting, where a slip often means the government simply pays more for the same end state.

Still, the audit makes clear that low contract growth should not be mistaken for a smooth program. Some of the underlying pressures have not disappeared. They have shifted into schedule compression, interface churn, and increasing demands on NASA’s own workforce.

SpaceX’s Lander Is the Immediate Schedule Problem

The audit’s bluntest statement is that SpaceX’s lander will not be ready for a June 2027 lunar landing. That judgment rests on a chain of milestones that continue to slide while technical work remains unfinished.

SpaceX’s Artemis III Starship development has already been delayed by at least two years since the 2021 award. A 15-month provider-requested delay moved the contractual delivery date from June 2025 to September 2026. NASA then pushed Artemis III to no later than June 2027. Even with that added time, the next essential milestone, a large-scale vehicle-to-vehicle cryogenic propellant transfer demonstration, slipped from March 2025 to March 2026.

That test is not just another engineering box to check. It is the foundation of the Starship lunar mission architecture. SpaceX plans to aggregate propellant in orbit using a depot and more than ten tanker flights before sending the lander onward. The company has demonstrated transfer between tanks within a vehicle, but not the full vehicle-to-vehicle operation the mission concept requires. The audit treats cryogenic fluid management as one of the provider’s hardest unsolved challenges, and that assessment reads as justified.

The deeper issue is sequence. Critical Design Review for Artemis III moved to August 2026. The uncrewed demonstration was expected to slip toward the end of 2026. That leaves only a narrow interval between demonstration and crewed use. In flight test terms, that is uncomfortable. In human lunar flight, it is severe.

Starship’s Development Style Brings Speed and Instability Together

The audit is not dismissive of SpaceX’s testing cadence. It notes meaningful progress. Since April 2023, the company conducted 11 integrated flight tests of Starship and the Super Heavy booster. There were successes, including in-space propellant transfer within the vehicle, controlled landings, and booster catch operations. There were also repeated failures. Flights 7, 8, and 9 ended with the loss of the Starship vehicle. The report notes that each mishap can add one to three months of schedule impact.

That is the trade embedded in SpaceX’s development culture. Fast iteration can surface problems quickly. It can also normalize a failure rhythm that becomes much harder to tolerate when a specific mission date and human-rating timeline are attached. For cargo launch systems, repeated test losses may be acceptable. For a human lunar lander that must integrate with Orion and the broader Artemis stack, they are far more consequential.

NASA’s response has been to request proposals from both SpaceX and Blue Origin on how to accelerate lander development for a 2028 Artemis III launch. The public record, as the audit presents it, does not yet allow a firm judgment on whether those proposals represent real engineering options or just better-crafted optimism. That part remains difficult to read with confidence.

Blue Origin Has More Time, Not Fewer Challenges

Blue Origin’s schedule picture is less immediate and less chaotic than SpaceX’s, but it is not comfortable. The Blue Moon lander for Artemis V slipped at least eight months from April 2028 to December 2028, and NASA later moved Artemis V to no later than March 2030. That extra time helps. It does not erase the technical concerns.

The audit says Blue Origin is still addressing propulsion maturity, mass reduction, and propellant margins identified at Preliminary Design Review. Nearly half of the official requests for action from that review remained open more than a year later. The agency expected Critical Design Review to move again, likely to July 2026. Its uncrewed demonstration mission, which is meant to prove cryogenic transfer capability, was then anticipated for February 2029.

Blue Origin’s mission architecture differs from SpaceX’s but depends on a similarly demanding set of in-space fueling operations. Blue Moon uses a transporter, refuelers, and transfers in multiple orbital regimes before reaching near-rectilinear halo orbit around the Moon. The technical phrase is cryogenic fluid management. The practical meaning is storing extremely cold propellants for long durations, limiting boiloff, transferring them reliably, and measuring them accurately in space. That is not routine work. It is frontier work.

The audit suggests Blue Origin benefits from a more phased block upgrade strategy and from learning opportunities before the crewed mission. That may prove true. Yet the report also makes plain that time on the calendar is not the same as maturity in hardware.

The Artemis Systems Are Tangled Together

The landers do not exist in isolation. They must work with Orion , Gateway , next-generation spacesuits, launch systems, and ground systems. The audit treats these as cross-program integration dependencies, and that phrase sounds bureaucratic until the examples arrive.

One example involved spacesuit interfaces. Changes in lander-spacesuit requirements added $26.2 million to the SpaceX contract in 2023. Blue Origin faced a different problem. It had begun work on a don and doff station in the airlock based on one government reference design, then learned the spacesuit provider was using a different connection approach. That forced Blue Origin toward redesign choices that could add both schedule strain and cost.

This is where Artemis starts to look less like a single program and more like a synchronized gamble among partially mature systems. Each element is waiting for the others to settle. That may be manageable in low Earth orbit. At the Moon, and especially at the lunar south pole, the margin for mismatched assumptions is much thinner.

NASA’s Insight Model Has Value, and It Is Expensive

The audit gives NASA substantial credit for its insight and collaboration model. Insight is not the same as formal approval. It is a system for monitoring provider work, reviewing data, performing independent analyses, and increasing visibility into higher-risk areas. Oversight remains the exercise of formal authority. Insight fills the broad territory between contractor autonomy and government ignorance.

By July 2025, NASA had insight into more than 1,100 focus areas across the two providers. More than 300 of those areas involved deeper levels of analysis. The report highlights engine development, cryogenic management, reaction control systems, and crew training as places where NASA increased attention because risk demanded it. That is a sign of a program trying to adapt rather than simply watch milestones slip.

The cost is real. NASA had spent more than half a billion dollars performing insight into the landers. As of mid-2025, 59 percent of the HLS civil service workforce was doing insight work across program offices. On top of that, NASA had provided nearly $66 million in equivalent personnel support through collaborations with the providers. Those collaborations included technical help on landing site selection, hazard detection, manual control design, and propellant gauging in low gravity.

The report’s tone here is measured. These efforts are useful. They also are not free. A commercial contract structure can keep provider invoices down while quietly pushing cost and labor demand back onto the agency.

Government Task Agreements Exposed a Simple Management Failure

Not every major finding in the audit is about rockets. One of the clearest administrative failures involved Government Task Agreements, which are the mechanisms that let providers use NASA facilities, assets, and services.

The problem was not that GTAs existed. The problem was that NASA lacked a formal process for handling agreements that had been proposed and then canceled, left unfulfilled, modified, or newly requested after contract award. That ambiguity led to confusion and a pause in GTA work in early 2025. After reconciliation, NASA authorized a $1.5 million decrement from SpaceX’s contract in 2026. No decrement was judged necessary for Blue Origin.

This is the sort of finding that can sound minor next to moon landers and life support systems. It is not minor. It shows that even when the top-level acquisition model is holding costs relatively steady, the supporting machinery of contract administration still has gaps. Those gaps do not make headlines until money has already moved and work has already stalled.

Safety Work Is Real, but the Audit Thinks It Is Incomplete

The report does not accuse NASA of ignoring safety. It points to real hazard reduction work, including required lunar ascent tests in the uncrewed demonstrations and focused efforts on plume-surface interaction. NASA has also used data from lunar lander missions such as Firefly Aerospace ’s Blue Ghost to improve modeling of lunar dust behavior.

Even so, the audit’s judgment is that important gaps remain. The most striking example involves Test Like You Fly principles. NASA required the uncrewed demonstrations to show precision landing, post-landing health telemetry, and ascent from the lunar surface. It did not require flight-like inclusion of some crew systems such as the environmental control and life support system, the airlock, or the elevator. That means the test vehicles will not match the mass and operating reality of the crewed versions. It also means some propellant aggregation demands will not be exercised end to end before astronauts depend on them.

The elevator point is especially telling for Starship. On paper, an elevator may look like a subsystem detail. On the Moon, with surface tilt, dust, and human movement in suits, it becomes a mission enabler. Leaving it out of the uncrewed lunar demonstration removes a chance to learn in the actual environment where the hardware must work.

Manual Control Is a Live Dispute, Not a Settled Requirement

Another safety issue in the audit is manual control. NASA and SpaceX disagree on whether SpaceX is meeting the intent of NASA’s manual control requirement, and Blue Origin had not yet made key decisions on its manual control design. That matters because manual control is not decorative redundancy. It is bound up with human-rating and with the last layer of crew survival if automation fails or behaves outside expected limits.

Here the audit is making a larger point about institutional memory. NASA has lived through decades of crewed flight experience in which backup modes and crew authority mattered. The report recommends that NASA consult lessons from the Commercial Crew Program’s manual control waiver process before HLS certification. That recommendation carries an implied warning. When a new spacecraft architecture departs too far from the habits shaped by earlier human spaceflight, the burden of proof should rise, not fall.

A fair reading of the audit is that NASA has not resolved this matter early enough.

Crew Survival Is Where the Report Turns Sharpest

The harshest section of the audit concerns crew survival analyses. NASA uses these studies to determine whether a crew can survive catastrophic events such as a fire or major system failure. If the analysis finds no available way to reach safety, then a crew survival gap exists.

The Inspector General argues that these analyses are arriving too late in the design cycle to influence hardware effectively. By the time the issue is identified, the vehicle may be too mature to change without major schedule impact or added expense. The analyses also look at single catastrophic events rather than simultaneous failures, and they do not account well for extended survival after the immediate danger has passed.

That last point is not abstract. A crew might survive the first phase of an emergency and still be stranded in a degraded state for days. The report says NASA’s present framework does not adequately model that condition. The recommendation to update survival analyses to include extended crew survival strategies is one of the most consequential in the entire document.

The audit also notes that the lander is the top contributor to loss-of-crew risk for Artemis III and IV. NASA’s threshold for lunar operations is 1 in 40, and 1 in 30 for Artemis missions overall. Current estimates show Starship meeting those thresholds. The report then undercuts any easy reassurance by invoking history. Early Apollo risk and early Space Shuttlerisk were both worse in practice than the programs believed at the time. That is the right place for skepticism.

NASA Has No Rescue Capability for Early Crewed Artemis Missions

The most objectiveing finding is also the simplest. NASA has ruled out crew rescue capabilities for the early crewed Artemis missions. If the lander becomes disabled on the lunar surface, or cannot dock with Orion or Gateway in lunar orbit, the crew would be lost.

NASA officials cited real obstacles. A rescue architecture would require extra consumables, rapid assembly and launch of rescue elements, and a safe haven for the crew while rescue is attempted. It would also be expensive, especially if it meant keeping spare Orion or spare landers in reserve.

Those objections are understandable. They are not fully persuasive. The stronger lesson from the audit is that the current Artemis architecture carries expeditionary risk without expeditionary recovery options. That may be tolerable as a temporary reality. It should not be normalized as acceptable design philosophy for sustained lunar operations.

The February 2026 Addendum Changes the Picture, but Not the Core Findings

After the audit fieldwork ended, NASA changed the Artemis campaign approach in February 2026. The new plan adds another Artemis mission in 2027 and recasts Artemis III as a low Earth orbit flight test rather than the first lunar surface landing. Under that concept, Orion would rendezvous and dock with one or both landers, and NASA would conduct integrated testing of life support, communications, propulsion, and possibly spacesuits before proceeding to an Artemis IV lunar landing in 2028.

This shift does not invalidate the audit. It reinforces it. NASA appears to have accepted, in operational terms, what the Inspector General had already concluded analytically: the original surface landing schedule was not aligned with the maturity of the landers and the integrated mission stack. Moving Artemis III into low Earth orbit is a sign that schedule realism is finally catching up with engineering reality.

Summary

This audit is not a story of runaway spending. It is a story of a program that has kept contract growth surprisingly controlled while accumulating enough technical, schedule, and safety strain to force a strategic reset. That distinction matters. Artemis is not failing in the most familiar Washington way. It is encountering a different kind of problem, where fixed-price discipline masks how much uncertainty has been pushed into timing, integration, and risk acceptance.

The document also makes a broader point about the future of lunar exploration. Commercial contracting can be effective for development. It does not reduce the government’s duty to understand what it is buying, to test it in representative conditions, and to plan for the day something goes wrong far from home. The next phase of Artemis will be judged less by whether NASA signed innovative contracts and more by whether it built a system in which astronauts have enough margin to survive both expected malfunctions and the unwelcome surprises that every new exploration era eventually brings.

Appendix: Top 10 Questions Answered in This Article

What is the Human Landing System in the Artemis program?

The Human Landing System is the lunar lander capability NASA needs to take astronauts from lunar orbit to the Moon’s surface and return them to orbit. In Artemis, that role is being developed by SpaceX for earlier missions and Blue Origin for a later sustaining mission.

Why did NASA use fixed-price contracts for lunar landers?

NASA used firm-fixed-price milestone contracts to shift cost control responsibility toward industry and pay for outcomes rather than ongoing labor. The audit found that this approach limited contract growth better than many traditional government development models.

Did the audit find major cost overruns in the HLS contracts?

No. The audit found relatively modest contract growth, with SpaceX up about 6 percent and Blue Origin up less than 1 percent as of December 2025. The larger problem identified was schedule delay rather than runaway contract cost.

Why is SpaceX’s Starship lander behind schedule for Artemis?

The report points to delayed cryogenic propellant transfer work, repeated flight test mishaps, tight milestone spacing, and the difficulty of proving an entirely new mission architecture. NASA concluded the lander would not be ready for a June 2027 lunar landing.

What is Blue Origin’s main challenge with Blue Moon?

Blue Origin faces major work in propulsion maturity, mass reduction, propellant margins, and long-duration cryogenic fluid management. Its architecture depends on reliable fuel storage and transfer across multiple orbital stages before lunar operations.

What does NASA mean by insight and oversight?

Insight is NASA’s monitoring and analytical access to contractor work, data, simulations, and tests. Oversight is formal government authority to review deliverables and concur or not concur with contractor actions.

What are Government Task Agreements and why did they matter in the audit?

Government Task Agreements allow contractors to use NASA facilities and services that are not otherwise available to them. The audit found NASA lacked clear policy for handling canceled, modified, or newly requested GTAs, which led to confusion and a contract cost reconciliation.

What safety concern did the audit raise about uncrewed demonstrations?

The audit said NASA was not fully following Test Like You Fly principles because the demonstration vehicles would not include some crew-related systems and would not fully represent crewed mission conditions. That reduces the amount of flight-like learning before astronauts depend on the landers.

Why is manual control such an important issue for HLS?

Manual control is part of the human-rating question because it provides a fallback if automated systems fail or behave unexpectedly. The audit found NASA and SpaceX disagreed on whether the requirement’s intent was being met, and Blue Origin had not finalized key design choices.

Does NASA have a rescue plan if astronauts are stranded during early Artemis lunar missions?

The audit says NASA has ruled out rescue capability for the early crewed Artemis missions. If a lander cannot leave the surface or cannot dock in lunar orbit, there is no committed rescue system ready to recover the crew.