Astronauts Return to ISS After Sheltering in SpaceX Dragon During Air Leak Repairs

At 9:04 a.m. EDT on Friday, June 5, 2026, NASA mission control in Houston issued an order that no astronaut wants to hear: get to your spacecraft and prepare to leave. Five crew members aboard the International Space Station scrambled into the SpaceX Crew Dragon "Freedom" capsule, sealed the hatch, and waited — ready to undock and plunge back to Earth if the air leak in the Russian segment of the station worsened beyond control [1][2].

Roughly two hours later, the order was reversed. Roscosmos had paused its repair work to take additional measurements, the leak rate stabilized, and the crew floated back into the station to resume operations [3]. But the brevity of the shelter event belies the severity of what it represents: the ISS, humanity's most expensive engineering project, is deteriorating faster than its operators can fix it.

What Happened on June 5

The immediate trigger was a repair attempt gone cautious. During cargo operations with the Progress 95 spacecraft earlier in the week, Roscosmos detected an increase in the air leak rate from the PrK transfer tunnel — a small vestibule connecting the Zvezda service module to an aft docking port — to approximately two pounds per day [4]. Cosmonauts identified two new suspected leak sites. They sealed one with Germetall-1, a two-component sealant that has become a familiar tool in Zvezda repairs, but the second remained unaddressed [5].

The planned next step was more aggressive: cosmonauts intended to use a saw to cut a structural bracket inside the PrK, providing better access to a suspected crack source for inspection. NASA determined that the procedure carried elevated risk to the surrounding structure and ordered its crew — the four members of SpaceX Crew-12 (NASA astronauts Jessica Meir and Jack Hathaway, ESA astronaut Sophie Adenot, and Roscosmos cosmonaut Andrey Fedyaev) plus NASA astronaut Chris Williams — into safe haven aboard the Dragon capsule [6][7].

After Roscosmos paused the cutting procedure and conducted additional measurements, NASA lifted the safe haven order. The crew returned to the station, and the PrK hatch was closed again per standing protocol [8].

Two Hours in a Capsule: How Close to the Edge?

The shelter period lasted approximately two hours — well within the Crew Dragon's operational margins. The capsule, designed to support four crew members for missions lasting weeks to months (SpaceX's Demo-2 test flight lasted nearly 64 days), carries its own life support, power, and environmental control systems [9]. Five crew members in a capsule rated for four would have been tight but manageable for the short duration.

The more relevant question is not how long the crew could have survived in the Dragon, but how quickly they could have escaped. In a safe haven configuration, the Dragon is sealed, powered, and capable of undocking within minutes. The crew keeps their flight suits accessible and the vehicle's systems active. If the leak had escalated to a point where station pressure was dropping uncontrollably, undocking and reentry would have been measured in minutes, not hours [10].

NASA has not publicly disclosed a specific pressure-loss-rate threshold that triggers mandatory evacuation versus shelter-in-place, though the agency has described the decision framework as based on real-time assessment of leak rate trends, structural risk, and crew safety margins [11].

A Leak Seven Years in the Making

The June 5 incident did not emerge from nowhere. The Zvezda module's PrK tunnel has been losing air since at least September 2019, making this one of the longest-running structural problems in ISS history [12].

Source: NASA / Roscosmos Reports

Data as of Jun 6, 2026CSV

The timeline tells a story of escalating severity. The leak began as a barely detectable pressure drop — less than 0.2 pounds of air per day. By August 2020, Roscosmos publicly acknowledged the problem. In 2021, Russian cosmonauts located and sealed two cracks, briefly reducing the loss rate. But the tunnel kept bleeding. By February 2024, the rate had climbed to 2.4 pounds per day. In April 2024, it hit 3.7 pounds per day — nearly twenty times the original rate — prompting NASA to classify the problem at the maximum 5×5 on its risk probability-and-consequence matrix [13][14].

The response since late 2024 has been to keep the PrK hatch closed except when a spacecraft is actively being accessed at the aft docking port, effectively quarantining the leaking section. This brought the effective loss rate down, and a repair campaign in June 2025 provided a temporary reprieve. But the new leaks discovered in early June 2026 demonstrate that the underlying structural problem remains unresolved [15][16].

The Root Cause Nobody Can Name

Despite seven years of investigation, the exact mechanism causing the cracks in the PrK remains unknown. Both NASA and Roscosmos suspect issues with welds, and engineering teams have investigated several hypotheses including pump vibration fatigue and environmental-assisted cracking — a phenomenon where the combination of stress and the space environment accelerates material degradation [17].

The Zvezda module itself was launched in July 2000, making it 26 years old — well past its original 15-year design life. It provides critical life support functions for the entire station, including oxygen generation, carbon dioxide removal, and water recycling. Its structural degradation is not merely an inconvenience; it is a threat to the habitability of the ISS as a whole [18].

Russian crews have deployed a range of repair techniques over the years: dust accumulation tests to locate air flow patterns that reveal crack locations, applications of Germetall-1 sealant, and physical patching. The June 2026 attempt to use a saw to access deeper structural areas represented an escalation in repair aggressiveness — and the fact that it was paused before completion suggests the engineering teams are operating at the boundary of what they can safely attempt without risking further damage [5][6].

The Financial Weight of an Aging Station

NASA spends approximately $3 billion per year operating the ISS, a figure that has climbed steadily as the station ages [19]. Maintenance and upgrade costs rose 35% between FY2016 and FY2020 according to a NASA Office of Inspector General report, and the trend has continued [20].

Source: NASA Budget Documents / Planetary Society

Data as of Jan 1, 2026CSV

Precise cost data for individual sheltering events is not publicly itemized, but the financial impact extends beyond the immediate emergency. Each unplanned shelter-in-place interrupts ongoing experiments — some of which involve biological samples or crystal growth processes that cannot be paused and restarted. Crew time diverted to emergency procedures and subsequent safety reviews represents hours that were budgeted for science. And consumables — air, water, food — used at higher rates during emergency configurations must be resupplied by cargo missions that cost tens of millions of dollars each [21].

The spacesuit maintenance program provides a telling parallel for how aging ISS systems strain budgets. A contract originally valued at $324 million over five years in 2010 had ballooned to $1.5 billion through 2027 as of July 2025, driven by supply chain challenges, parts obsolescence, and quality issues with hardware designed in the 1970s [22].

Who Is Aboard — and Who Answers to Whom in an Emergency

At the time of the June 5 incident, the ISS hosted a multinational crew. The Crew-12 contingent aboard the Dragon Freedom included NASA's Jessica Meir and Jack Hathaway, ESA's Sophie Adenot (French), and Roscosmos's Andrey Fedyaev (Russian). NASA astronaut Chris Williams, who arrived via a Russian Soyuz, joined them in the Dragon. On the Russian segment, cosmonauts continued repair work [6][23].

The emergency highlighted a structural peculiarity of ISS operations: NASA ordered its crew members into the American spacecraft while Russian cosmonauts remained in the Russian segment to conduct the repair. In a full evacuation scenario, crew members return to the spacecraft they arrived in — Dragon for Crew-12 and the Soyuz for those who launched aboard it. This means emergency escape is governed by arrival vehicle, not nationality, though the practical effect is that American and allied astronauts head to one capsule while Russian crew members head to another [10].

Joint US-Russian emergency protocols exist and are exercised periodically, but the specifics of when they were last jointly tested or updated have not been publicly detailed in the context of the current geopolitical environment. The broader ISS partnership agreement, renegotiated in 2022, extends cross-flight arrangements through at least 2025, and operational cooperation has continued at the crew level even as diplomatic relations between Washington and Moscow have remained strained [24].

The Skeptic's Case: Is It Time to Leave?

The June 5 shelter event was at least the third publicly reported instance in recent years where ISS crew members were ordered to prepare for possible evacuation due to structural concerns. Critics of continued ISS investment argue that the pattern is unmistakable: an aging station is producing emergencies at an increasing rate, and the $3 billion annual operating cost would be better directed toward commercial successors [25].

The NASA Office of Inspector General has warned that "concerns about the long-term viability of some existing International Space Station modules and the potential of delays in development of commercial space stations heighten the risk of a gap in low Earth orbit destinations" [20]. The OIG has also noted that research needed for long-duration missions to the Moon and Mars — one of the primary justifications for ISS operations — will not be complete by the station's planned 2030 retirement date [26].

Academic interest in space station structural integrity has surged in parallel with the station's problems. Research publications on the topic peaked at 8,538 papers in 2025, up from 1,615 in 2011 — a more than fivefold increase reflecting growing scientific concern about orbital infrastructure aging [27].

Source: OpenAlex

Data as of Jan 1, 2026CSV

The Proponent's Case: There Is No Alternative — Yet

ISS defenders counter with a blunt practical argument: there is nowhere else to go. None of the commercial stations under development — Axiom Station, Orbital Reef (Blue Origin and Sierra Space), Starlab (Voyager Space and Airbus), or Vast's Haven-1 — are expected to reach full operational capability before the early 2030s at the earliest [28].

Axiom Space plans to attach its first module to the ISS itself before eventually separating into a free-flying station, potentially by 2028. Vast targets a Q1 2027 launch for its Haven-1 single-module station. Starlab targets 2029. But the OIG has flagged "limited market demand, inadequate funding, unreliable cost estimates, and still-evolving requirements" as obstacles that could push these timelines further right [26][29].

A premature ISS retirement would create what the OIG calls a gap in low Earth orbit access — halting microgravity research, threatening the commercial space economy that has grown around ISS operations, and delaying preparations for Artemis lunar missions that depend on ISS-based health research [20]. A Senate bill under consideration would extend ISS operations to 2032, adding two years beyond NASA's current 2030 target [17].

The safety argument cuts both ways. Yes, the station is aging and producing more emergencies. But each emergency also demonstrates that existing protocols work: the crew sheltered successfully, the leak was managed, and operations resumed within hours. Proponents argue that the station's safety systems, including redundant life support, multiple docked escape vehicles, and compartmentalized design that allows problem areas to be isolated, remain adequate for continued operation even with known degradation [3][11].

What Comes Next

Roscosmos has not yet completed repairs to the second leak site identified in the PrK. Additional measurements and structural assessments are ongoing. The PrK hatch remains closed except during active spacecraft operations at the aft port — a workaround, not a fix [8][15].

Joint NASA-Roscosmos commission meetings in Houston have not yet produced consensus on the severity of the Zvezda cracks or the module's remaining safe operational life. Partner agencies are pushing for a final decision on station deorbiting or further extension by the end of 2026 [17].

The fundamental tension is straightforward: the ISS is too important to abandon and too old to trust unconditionally. Every patch of Germetall-1 sealant buys time, but no amount of sealant can reverse a quarter-century of thermal cycling, micrometeorite impacts, and material fatigue in the vacuum of space. The question is no longer whether the ISS will be retired, but whether its successors will be ready when it must be.