NTSB: Amazon 767 Crew Briefly Attempted Go-Around Before Aborting, Thrust Reversers and Speed Brakes Never Deployed

The National Transportation Safety Board released flight-data findings from the Boeing 767-300F crash at Miami International Airport that killed five people and injured five others after the freighter overran Runway 30. According to the flight-data recorder timeline, the nose landing gear and right main landing gear touched down at 158 knots ground speed with 30 seconds remaining on the recording, and wheel brakes were applied seven seconds later at 146 knots. The left main landing gear did not touch down until 19 seconds before the end of the recording — 11 seconds after initial touchdown — at 134 knots.

At 15 seconds before the recording ended, with the aircraft at 120 knots, the wheel brakes were released and engine power was increased, consistent with the crew initiating a go-around. Four seconds later, at 117 knots, the throttles were pulled back to idle and the brakes were reapplied, indicating the pilots aborted that attempt. At seven seconds before the end of the recording, the aircraft's movement changed in several lateral, longitudinal, and vertical directions at 96 knots, and the recording ended with the aircraft travelling at 65 knots. The NTSB confirmed that data shows no indication thrust reversers or speed brakes were deployed at any point during the landing, and investigators noted that the 11-second delay before the left main gear touched down likely prevented those systems, which require the aircraft to be firmly on the ground, from activating.

Runway 12/30, the runway used, has the shortest landing distance available of Miami International Airport's four runways at less than 8,000 feet, with approximately 1,000 feet of runway safety area beyond the pavement that does not feature an Engineered Materials Arresting System. Both pilots survived and were treated for injuries at the scene before being released. NTSB investigator in charge Chihoon Shin said investigators need to speak with the crew to understand how they were manipulating the controls, and both pilots were scheduled to be interviewed by the NTSB on Wednesday.

Carroll had accumulated 7,145 flight hours as of March and received his Boeing 767 type rating only in May, four months before the accident. Silva had 2,655 flight hours as of August and had been rated on the 767 since April 2025. Both pilots were on their third flight of the day, having operated sectors from Cincinnati to Miami and Miami to San Juan before the accident flight. The two pilots were employed by 21 Air, which began operating Boeing 767 freighter services on behalf of Amazon in late 2024, and the airline has faced safety allegations from former employees, including claims that pilots were discouraged from submitting safety reports and that important maintenance work was regularly deferred.

This sequence belongs to a category investigators watch closely in overrun cases: a delayed second main-gear touchdown that leaves the aircraft in an ambiguous air-ground state just as the crew has to decide between stopping and going around. Thrust reversers and speed brakes on transport aircraft are typically gated by weight-on-wheels logic, so an 11-second gap between the first and second main gear making contact can silently remove two of the primary stopping systems from a landing roll that still looks, to the crew, like it is proceeding normally.

Runway safety areas exist precisely for the moments when deceleration falls short of what the runway length assumed, and the absence of an Engineered Materials Arresting System on this stretch of pavement means any margin beyond the roughly 1,000 feet available had to come from the aircraft's own systems working correctly. The abandoned go-around adds a second variable: once thrust was reapplied and then reduced again within seconds, the aircraft was neither fully committed to stopping nor to flying, a transition point NTSB will need the pilots' own account to reconstruct.

The interviews scheduled for both pilots, together with any findings on 21 Air's training and maintenance practices given the prior safety allegations, are the next concrete markers. Whether the crew's control inputs during the touchdown and go-around sequence align with the flight-data timeline will shape how regulators and Amazon's freighter contractors assess pilot experience thresholds on newly type-rated crews flying multi-sector days.