Amazon Doesn't Fly Planes: The Contractual Machine Behind the Miami Crash

A Boeing 767 cargo aircraft operated by 21 Air nosed 1,300 feet past the end of a Miami runway on September 6, hit two vehicles and killed five people. All five were employees of a ground-cleaning company. Amazon's stock barely moved, as though it had been reminded of its scale.

The more interesting question isn't whether the crash is tragic — it is, and it shouldn't be treated as a mere business event — but whether it actually changes anything about Amazon's business. The answer depends on understanding what Amazon actually built in the sky, and who it designed to take the risk.

Amazon does not fly planes. Not really.

The company has spent a decade building Amazon Air, a cargo network that supports the next-day and two-day deliveries that define the Prime promise. But Amazon Air isn't an airline. It's a virtual airline — a brand and a scheduling system wrapped around a fleet of more than 100 aircraft across nine separate contracted carriers who hold their own FAA operating certificates. 21 Air, the carrier behind the Miami crash, was one of them, flying eight planes under the Amazon brand.

Here's the plumbing. Under Amazon's arrangement, the carrier — 21 Air — provides the crew, the maintenance, and the insurance. That bundle is called a CMI lease, and it matters because the Federal Aviation Administration looks at the operating certificate to determine who's responsible for safety. The operating certificate says 21 Air was flying the plane. That means 21 Air's safety protocols, 21 Air's pilot training, 21 Air's insurance. Amazon doesn't hold the certificate. Amazon doesn't employ the pilots. Amazon doesn't own the crashed aircraft — it was leased from Titan Aviation, a sister company of Atlas Air, Amazon's former primary cargo partner.

This is the same contractor-separation strategy Amazon uses on the ground with its Delivery Service Partners. Those independent business owners wear Amazon uniforms, drive Amazon-branded vans, deliver Amazon packages, and follow Amazon's rules. But Amazon insists they aren't Amazon employees. The legal line has come under pressure there — a Georgia jury awarded $16.2 million in a crash case, finding Amazon 85% responsible, rejecting the contractor shield. The aviation side has historically held up better, because the FAA framework gives that operating certificate real legal teeth.

The official description of Amazon Air is a logistics network. In practice, it's closer to what happens when you buy the car, hire a driver, and let the driver's employer — not you — carry the operational and safety responsibility if something goes wrong.

That said, a wrongful death lawsuit has been filed naming Amazon, 21 Air, and several leasing entities. Aviation attorneys say Amazon is likely shielded from the main liability, since "operational control" falls on the carrier under federal law. Shielded doesn't mean immune — lawsuits named Amazon anyway, and the NTSB has said it will examine the contract between Amazon and 21 Air to determine what safety provisions are in place. The investigation is the real question mark, not the litigation itself. Most aviation claims are settled privately through insurance, and if 21 Air carries adequate coverage (which cargo carriers typically do), the financial exposure for any defendant with Amazon's balance sheet is a rounding error.

But let's stay with the mechanism for a moment, because there's a second layer to this story that's less visible and more structurally interesting.

Amazon didn't just contract out the flying to avoid liability. It structured the deal so the carriers carry the risk of underutilization too. In white-label arrangements like 21 Air's, the customer — Amazon — guarantees payment for a minimum number of flight hours regardless of load factors. That means Amazon absorbs the capacity cost, but the carrier runs the operation and holds the safety certificate. Amazon gets predictable capacity and the carrier gets a guaranteed revenue floor. It's a split that lets Amazon scale its delivery network without building an airline — without a pilot union, without an FAA Part 121 certification, without the operational overhead of actually running a carrier.

There's an incentive tension here that the crash is now forcing into public view. Amazon wants reliability and safety, obviously. But the carriers competing for these contracts compete partly on price, and 21 Air's former pilot union president filed a whistleblower lawsuit in 2022 alleging the company had safety issues long before the crash, and a departing safety director's 2021 email accused leaders of paying "lip service" to safety. He said the company struggled to attract experienced pilots because it paid less than UPS or FedEx. The captain on the Miami flight had been certified on the Boeing 767 for less than five months; the first officer, less than 18 months. A retired airline captain who reviewed the cockpit recordings told NPR he'd never seen "such a terrible performance from a crew" in 60 years of flying.

Amazon says the safety allegations predate its current relationship with 21 Air and that it only contracts with FAA-certified carriers — which is true, but that's a low bar. An operating certificate means you've met minimum standards, not that your training culture is strong. This is the gap between certification and competence that the Miami crash exposed.

For investors, the practical consequence is narrow. Amazon's stock is trading at $256.78, up 1.9% today and 11.2% year-to-date, on a company with a $2.77 trillion market cap. The crash itself is not a financial event for Amazon. The pause with 21 Air affects the eight planes it operated for Amazon, a small share of the fleet Amazon contracts across its carriers, and Amazon can redistribute those routes among its remaining carriers. The logistics disruption is real for a few days but doesn't change the delivery economics that drive Prime retention.

What could change is the regulatory or reputational arc. If the NTSB finds that Amazon's contracting structure systematically incentivizes carriers to cut corners on pilot experience and training, that opens a different kind of risk. Not a financial one — more of a structural one. The same contractor model that insulates Amazon from liability could become a liability if regulators decide the insulation itself is the problem. That's what happened with the DSP model on the ground, where juries started piercing the contractor veil.

The Miami crash is the third major accident involving a Boeing 767 flying for Amazon — a 2019 crash near Houston that killed three people, and a 2024 runway overrun in Vancouver. Atlas Air, the operator in the 2019 crash, was dropped from Amazon's domestic network partly in response. The pattern suggests Amazon responds to catastrophic events by swapping out carriers rather than rethinking the structure that produced them. That's rational in the short run. It may not be in the long run if the structure itself creates the risk.

None of this is a reason to buy or sell Amazon stock today. The company's logistics costs are a fraction of its $717 billion in 2025 revenue, and the air network — for all its operational complexity — is a cost center designed to make Prime feel worth it. The crash is a reminder that Amazon's growth is built on a chain of contractual relationships where each link is designed to push risk to the next participant. That's not unusual for a company of this size. It's just worth understanding what you actually own, and where the risk is supposed to live, before a headline makes it sound like it's all sitting on one balance sheet.