Flying to the Moon for the First Time in 54 Years Is Risky — But How Risky?

Lead

NASA’s Artemis II — the first crewed mission to the Moon’s vicinity since 1972 — is proceeding toward a planned liftoff no earlier than April 1, 2026 at 6:24 p.m. EDT (22:24 UTC) from Kennedy Space Center. The four-person crew will ride the Space Launch System (SLS) rocket and Orion spacecraft on a roughly nine-day circumlunar test that will travel thousands of miles beyond the Moon’s far side. Agency leaders stress the flight is a test and acknowledge significant unknowns, citing limited flight data (only Artemis I in 2022) and recent ground issues with fueling hardware. Mission managers approved continuing final preparations after a flight readiness review, while declining to publish a single definitive probability for catastrophic failure.

Key takeaways

• Artemis II is the first human mission beyond low-Earth orbit since 1972 and will last about nine days from launch to splashdown, carrying four crew members aboard SLS and Orion.
• NASA scheduled liftoff no earlier than April 1, 2026 at 6:24 p.m. EDT (22:24 UTC) and added April 2 among six early-April launch opportunities; further opportunities run later in April.
• Ground work encountered a leaky hydrogen seal and a helium-loading issue that required returning the rocket to the hangar; seals were replaced and managers chose not to perform another full fueling test before the launch attempt.
• Agency leaders publicly questioned the usefulness of a single probabilistic failure number for Artemis II; mission leadership says they aim to reduce loss-of-crew odds toward 1-in-50 but acknowledged the mission likely does not meet that target.
• Top technical risks cited include micrometeoroids and orbital debris (MMOD) and Orion’s environmental control and life support system (ECLSS); the Launch Abort System reduces some ascent risk for crew escape.
• Historical context matters: past agency assessments (Artemis I at 1-in-125 for loss of the Orion vehicle) and earlier programs show probabilistic figures can be misleading early in a program when data are sparse.

Background

The Artemis program aims to return astronauts to the lunar neighborhood and eventually to a surface landing near the Moon’s south pole. After Artemis I, an uncrewed lunar test in 2022, Artemis II will be the first crewed circumlunar mission in 54 years, exercising deep-space operations, navigation, and reentry at speeds near Mach 39. Artemis architecture couples the heavy-lift SLS rocket and the Orion crew vehicle; those two elements have flown together only once (Artemis I), leaving program managers with a single integrated-flight data point.

Human spaceflight risk assessment historically blends engineering analysis, test data, and judgment. When a program has little flight history, probabilistic risk models rely heavily on assumptions and expert inputs; that makes early numbers volatile and sometimes deceptive. The long gap between missions raises operational concerns: teams lose launch cadence, skilled personnel rotate off programs, and subtleties in hardware and procedures can change between flights, complicating direct comparisons to more mature launchers.

Main event

NASA held a two-day flight readiness review at Kennedy Space Center where program leads described the agency’s “risk posture” and the mitigations they had in place. Acting associate administrator Lori Glaze said the review was open and transparent but questioned whether a single numerical failure probability tells the whole story. Managers unanimously approved moving ahead with final launch preparations, with the next rollout to the pad scheduled after completing repairs.

Earlier troubleshooting prompted a rollback: engineers found a leaking hydrogen seal in the SLS fueling line and later encountered a helium-loading anomaly in the rocket’s upper stage, forcing the vehicle back to the hangar for inspection and repair. After replacing seals and completing a full countdown rehearsal that showed no leaks, managers decided not to repeat a full wet dress rehearsal before attempting a launch. According to mission leadership, the next time the SLS will be fully tanked is when it’s attempting liftoff.

Mission manager Matt Ramsey said MMOD and Orion’s environmental control and life support system were among his principal operational concerns; the ECLSS had not flown with full capability on Artemis I. John Honeycutt, chair of the Artemis II mission management team, gave a blunt assessment of the limits of probabilistic numbers and emphasized focusing on the technical failure chains that produce catastrophe rather than over-relying on a single global probability.

Analysis & implications

Probabilistic risk assessments (PRAs) aim to quantify likelihoods by aggregating failure modes, redundancy, and common-cause events, but PRAs are only as reliable as their inputs. With a single integrated flight (Artemis I) and limited operational repetitions, uncertainty margins remain large. NASA officials cautioned that numerical outputs can mislead decision-makers and the public when underlying assumptions are poorly constrained.

Operational cadence is a practical safety factor. Longer gaps between flights reduce the opportunity to learn iteratively from real missions; workforce turnover and deferred hardware refinements can raise programmatic risk even if individual component reliabilities remain high. That dynamic helps explain leadership’s push to shorten the interval to future Artemis missions while balancing technical readiness and lander development schedules.

Program-level targets (Honeycutt cited a goal of getting to less than 1-in-50 chance of crew loss) provide useful objectives, but setting them without sufficient validation risks overstating certainty. The Office of Inspector General’s published loss-of-crew threshold of 1-in-30 for Artemis adds a governance anchor, but managers have not released a finalized Artemis II probabilistic number. For policymakers and the public, transparency about assumptions and limits in PRAs matters as much as headline figures.

Comparison & data

These reported numbers illustrate how agencies publish probabilistic estimates that are shaped by the aircraft/vehicle maturity and the assumptions behind failure rates. Artemis II’s leadership has chosen not to publish a single consolidated probability for the flight, arguing that the mission’s novelty and sparse integrated-flight data make any single percentage brittle.

Reactions & quotes

Agency leaders, mission managers, astronauts and oversight bodies all framed the discussion differently; the public exchange at the flight readiness review highlighted those distinctions.

“Numbers have value but they don’t tell every story—we need to use them relatively and think about failure chains,”

Lori Glaze, NASA acting associate administrator (context: emphasized limits of single-number risk metrics at the Flight Readiness Review)

Glaze pressed for balanced use of PRAs while describing the specific technical areas the crew and managers reviewed during the FRR, such as entry interface targeting, communications, and life-support performance.

“We’re trying to think of everything that could go wrong and work through the demise chains rather than hang our hat on a single probability,”

John Honeycutt, Artemis II mission management team chair (context: described the program’s focus on technical failure scenarios and historical lessons)

Honeycutt pointed to historical surprises—such as shuttle debris risks—when arguing for a systems-level, scenario-driven approach rather than overconfidence in an early PRA.

“I told my family about what to expect because this is an honest part of our job; we know there’s risk but we’re going to learn on this flight,”

Reid Wiseman, Artemis II commander (context: crew perspective on family conversations and personal readiness)

Wiseman framed acceptance of risk as a personal and professional reality for test astronauts while expressing full commitment to the mission.

Unconfirmed

• NASA has not released a finalized consolidated probabilistic number for Artemis II; reported targets (e.g., sub-1-in-50) reflect program goals rather than a published mission-wide PRA.
• The long-term effect of skipping an additional full fueling test on launch-day risk remains to be validated during the upcoming rollout and tanking; managers cite successful seal replacement and countdown rehearsal but full flight fueling has not yet been demonstrated.
• Timelines for Artemis III and a potential 2028 Artemis IV human landing hinge on commercial lander readiness and are contingent on near-flawless execution of Artemis II and subsequent missions.

Bottom line

Artemis II is a high-consequence test with both well-understood technical hazards (MMOD, ECLSS) and less-quantifiable programmatic uncertainties because the integrated vehicle has only flown together once. NASA’s leaders elected to press ahead after an open review that highlighted known risks, recent ground repairs, and the limits of probabilistic models when data are sparse.

For the public and policymakers, the most reliable picture of safety will come from transparent disclosure of assumptions, continued demonstration of hardware under flight-like conditions, and an honest accounting of what is known versus what remains uncertain. If Artemis II succeeds, it will build operational knowledge rapidly; if it encounters trouble, the investigation and subsequent adjustments will be essential to making future lunar missions safer.

Sources

• Ars Technica — reporting and interview coverage of the Artemis II flight readiness review (journalism)
• NASA Artemis program — official program information and press materials (official)