MH370 Was Reanalyzed by AI — And the Findings Shocked Everyone

Overview (2026 AI reanalysis of MH370)

In 2026, a team reanalyzed the disappearance of Malaysia Airlines Flight MH370 by feeding all known evidence into an AI system (“Grock,” built by X AI). The goal was to avoid the piecemeal guessing that persisted for more than a decade.

What the AI re-checking found (core conclusions)

• MH370’s disappearance wasn’t a simple accident scenario. Rather than treating the loss of communications as the beginning of an emergency, the AI aligned every data stream by time and identified a deliberate sequence in which tracking systems went dark one after another.

• A normal early flight behavior, followed by a sudden change. After takeoff from Kuala Lumpur, the aircraft initially behaved normally. After a routine radio exchange, the transponder shut off, and the plane performed a large, controlled-looking turn across multiple airspace boundaries. The AI ties this to manual inputs described in the 2018 Malaysian safety report, arguing the timeline suggests an intentional effort to stop being tracked.

• Two communication/tracking systems fail (or are disabled) in a critical window. Grock’s reconstruction suggests that two independent onboard communication/tracking systems failed or were disabled around the same time—consistent with planned silencing, rather than a single mechanical breakdown.

Why the Southern Indian Ocean search area became “locked in”

A key pillar of the new analysis is that, even after most tracking went offline, MH370’s satellite unit kept responding to routine pings from a ground network operated by Inmarsat.

The AI used the handshake signals in two main ways:

1. Timing differences between ping and response to generate possible position arcs.
2. Doppler shift (frequency changes) to narrow motion and direction.

This approach was used to infer that the plane was on a steady southern track for hours after radar contact ended, converging on a specific region of the southern Indian Ocean—described as the same “seventh ark” area later emphasized in official search efforts.

Evidence of how the plane ended

The AI claims the satellite log contains clues consistent with a final phase:

• After the last major arc, the satellite unit appears to “restart” (interpreted as backup power after a power drop).
• Fuel-exhaust timing is argued to match an interval when the plane would lose main engine power.
• The final satellite signals are interpreted as describing a steep, accelerating descent—far faster than a normal passenger descent—supporting a high-speed plunge into the ocean.

Physical debris consistent with a “high-speed, not-dithering” scenario

Debris recovery is presented as reinforcement for the AI and satellite findings:

• The Reunion flaperon and other wing components were analyzed. A key claim is that the flaperon tore off in a position consistent with high-speed flight (with flaps stowed) rather than configurations used to slow down for a gentle ditching.

• A separate study of barnacles on the flaperon (including work published in 2023) is described as an independent “geolocation” method. Barnacle shell growth layers can preserve seawater temperature histories; matching these temperatures to drift simulations purportedly points back toward the same southern Indian Ocean region indicated by the satellite arcs.

The “captain’s simulator” link (direction of suspicion, not identity)

• Malaysian investigators seized a flight simulator used by Captain Zahari Ahmed Shah.
• The FBI later recovered deleted data from it.

The AI/coverage claims the simulator contained a saved route (created weeks before MH370’s disappearance) that aligns strongly with the reconstructed real-world path—particularly the turn south and the fuel-exhaust ending over open ocean.

However, the video emphasizes important limitations:

A saved simulation is not proof of a confession. It remains unclear who saved the files or whether the same person was at the controls.

Attempted confirmation via Ocean Infinity deep-sea search

• Between December 2025 and January 2026, Ocean Infinity used deep-sea robotics to search the specific high-probability seabed area implied by the “seventh ark.”
• The search covered thousands of square kilometers, but found no wreckage—leaving families still without “where” and “why.”

Overall argument of the coverage

The central thesis is that multiple independent evidence streams converge on a scenario in which the aircraft was intentionally moved into a region where recovery was unlikely—by disabling tracking and exploiting blind spots—followed by a violent high-speed water impact that caused wreckage to sink rapidly.

The AI is described as able to reconstruct the route and mechanics of the final moments, but it cannot identify the responsible person or motive, and the ocean search has not produced final physical confirmation.

Presenters or contributors

• Narrator / Video host (not named in the subtitles)
• Grock / AI team (described as built by X AI)
• Peter Foley (Search Director, ATSB)
• Gregory Herbert (University of South Florida, biologist)
• Britain’s DST (Defense Science and Technology Laboratory) / analysts (referenced)
• Ocean Infinity (deep-sea robotics company; company team not individually named)
• Captain Zahari Ahmed Shah (subject of the referenced simulator data)
• ATSB (organization referenced; other specific staff not named)
• Inmarsat (company operating the satellite ping service; not an individual)