All three hydraulic systems on an Air India Airbus A320 registered low pressure within four seconds at 36,000 feet, and recovery began roughly nine seconds after the first warning, according to a preliminary investigation released Friday. The August 4 upset on flight AI2379 from Phuket to New Delhi injured 20 passengers and four cabin crew members after the aircraft abruptly climbed and descended while the seatbelt signs were off. India’s Aircraft Accident Investigation Bureau has reclassified the event from a serious incident to an accident.
The report establishes an unusually compressed sequence but not its cause. It does not determine whether the aircraft suffered a physical loss of hydraulic pressure, received erroneous indications or encountered another failure. It also records that the captain later returned a “non-negative” confirmatory test for a psychoactive substance, without naming the substance or linking that result to the technical event.
That distinction matters for travelers and for aviation oversight. A preliminary report preserves evidence and identifies immediate safety issues; it is not a finding of blame. The most consequential questions remain open: what produced three near-simultaneous low-pressure indications, why primary flight controls were briefly unavailable, how the crew response affected the aircraft’s path and whether any maintenance, component or procedural change is required.
Nine Seconds From Warning to Recovery
Flight data put the aircraft in level cruise when the green hydraulic system first indicated low pressure. Blue and yellow followed within four seconds. The autopilot disconnected, a continuous cockpit warning sounded and a stall warning lasted about two seconds. The aircraft climbed 372 feet above its assigned altitude and then fell 292 feet below it before the crew stabilized the flight.
Hydraulic power moves critical control surfaces and other aircraft systems. The A320 uses three independent circuits—green, blue and yellow—to provide redundancy, so a brief loss across all three is materially different from a fault isolated to one circuit. An earlier Airbus assessment reviewed by Reuters said elevators and ailerons were unavailable for about four seconds. The first officer’s nose-down command did not produce a direct response until hydraulic indications began returning.
The blue system recovered first, about nine seconds after the sequence began, with yellow and green returning seconds later. That restoration allowed normal control to resume, but it does not explain the initiating event. The Indian Express reported that investigators are still testing the central alternatives: a genuine loss of pressure, misleading sensor information, an electrical or wiring problem, or another mechanism not yet identified.
The Cabin Absorbed the Upset
There were 137 passengers, including three infants, and eight crew members aboard. The seatbelt signs were off during cruise when the aircraft moved sharply, and unrestrained occupants struck overhead bins and cabin surfaces. The interior was damaged, and 24 people needed medical assistance after arrival. Initial accounts described the event as turbulence, but the preliminary findings instead center on the hydraulic indications and resulting control interruption.
The pilots continued to Delhi after the aircraft stabilized and requested priority because medical help was needed. The flight landed safely on runway 11L. The preliminary record does not criticize that decision or conclude that diversion would have produced a better outcome. It does, however, turn a short technical interruption into a broader cabin-safety case: even a flight that remains controllable and lands normally can injure people when an unexpected upset occurs without warning.
An Associated Press account soon after the event reported fewer injuries because information was still developing. The preliminary report’s 24-person total supersedes that early count. The change illustrates why immediate airline statements, hospital tallies and formal investigation records can differ—and why a current assessment should privilege the newest official evidence.
The Technical Cause Remains Open
Investigators have collected components and samples and are examining flight-data and cockpit-voice recordings, maintenance history, airline and airport records, air-traffic-control material and interviews. The inquiry also covers medical and human-performance factors. India’s public broadcaster said the AAIB warned against conclusions drawn from incomplete or isolated information when it opened the case.
France’s Bureau of Enquiry and Analysis is participating because France is the aircraft’s state of design, and Airbus is providing technical assistance. That structure follows the international system for accident investigations, which assigns roles to the states connected to an aircraft, operator, design and manufacture. New ICAO standards adopted in March reinforced the principle that safety investigations must remain independent and should not be compromised by conflicts of interest.
So far, investigators have issued no safety recommendation to Airbus. That is significant but not exculpatory: it means the preliminary evidence has not supported an immediate fleet action, not that the aircraft, its components or maintenance can be ruled out. The final analysis must reconcile recorded warnings with physical evidence and determine whether the same combination could recur before regulators can judge whether inspections, software changes, component replacements or revised procedures are warranted.
A Separate Pilot Test Complicates the Case
The first officer, 32, was flying when the upset began; the captain, 34, later took control. Both were tested after the flight. The first officer tested negative, while the captain’s initial screen and confirmatory laboratory test were non-negative for a psychoactive substance. Air India said Friday that it terminated the captain after the confirmed result and had begun a one-time voluntary testing program covering all its pilots.
The test result is a serious compliance issue, but the published evidence does not show it caused the hydraulic indications, control interruption or altitude change. Nor does the report specify the detected substance, dose, timing or evidence of impairment. International guidance treats toxicology as one part of a larger inquiry: a test finding can require employment or regulatory action while causation still depends on operational, medical and technical evidence. Conflating the two would outrun the investigation.
The AAIB recommended that the Directorate General of Civil Aviation take appropriate action to prevent psychoactive-substance misuse among aviation personnel. India’s current framework includes random testing, and authorities are considering whether to require annual testing rather than sampling a portion of covered personnel. The DGCA’s published rules make testing a systemic oversight question, separate from the still-unresolved engineering chain on AI2379.
A Wider Test of Safety Confidence
Air India is under exceptional scrutiny after the June 2025 crash of flight AI171 in Ahmedabad killed 260 people, including 241 of the 242 aboard and 19 on the ground. That Boeing 787 disaster and this A320 accident involve different aircraft, circumstances and evidence; treating them as one technical pattern would be misleading. Their connection is institutional: the airline, regulators and investigators must demonstrate that evidence is preserved, disclosed and converted into corrective action.
Commercial flying remains statistically very safe. ICAO’s latest global review put the 2025 accident rate at 2.2 per million departures, down from 2.6 in 2024. Those figures do not diminish the injuries on AI2379; they show why investigations focus intensely on rare combinations that defeat redundant systems, even briefly.
The next useful evidence will not be another theory. It will be the investigators’ reconciliation of recorder data, hydraulic hardware, sensors, wiring, maintenance records and crew actions, followed by any findings on cabin restraint and medical factors. Until then, the firm conclusion is deliberately narrow: three hydraulic systems indicated pressure loss almost together, flight controls were briefly unavailable, the aircraft recovered and landed, and 24 people were injured. The mechanism—and any responsibility for it—remains unproven.
