Safety-first company. Cost-first execution.

The pressure

For most of the twentieth century, Boeing was the gold standard in commercial aviation. Engineers ran the company. Then came the 1997 merger with McDonnell Douglas — technically an acquisition by Boeing, but in practice a cultural takeover. The new leadership was explicit: Boeing needed to behave "less like an engineering firm and more like a business." Headquarters moved from Seattle, where the planes were built, to Chicago, where the money was managed.

In 2011, Airbus launched the A320neo with 15% better fuel efficiency. American Airlines, a Boeing loyalist, signaled it would switch. Boeing faced a choice: design a clean-sheet replacement for the 737 (7+ years, $15B+) or bolt new engines onto a 1960s airframe and deliver in 6. The board chose speed. Every decision that followed — MCAS, single-sensor architecture, minimal pilot disclosure — was downstream of that one commitment. Not downstream of malice. Downstream of schedule.

"Boeing was an engineering company that got bought by its own merger partner's spreadsheet culture."

The shortcut

The new engines did not fit under the 737's low-slung wings. Mounted forward and higher, they changed the aircraft's pitch characteristics — the nose tended to rise, risking aerodynamic stall. The solution was MCAS: software that automatically pushed the nose down when sensors detected a high angle of attack. Software to mask a hardware compromise.

MCAS had three fatal design flaws. It relied on a single angle-of-attack sensor — no redundancy. It could activate repeatedly, overriding pilot inputs with increasing authority. And Boeing did not prominently disclose it in pilot training materials. Airlines were told their pilots could transition to the MAX with an iPad course. No simulator time required.

Engineers raised the single-sensor risk internally. One test pilot wrote that MCAS was "running rampant" in simulator tests. Management classified these concerns through review processes that had been quietly reoriented around certification speed rather than safety margins.

"The system Boeing built to catch safety problems had been rewired to optimize for certification speed."

The cost

On October 29, 2018, Lion Air Flight 610 crashed into the Java Sea. All 189 on board died. MCAS had fired repeatedly on faulty sensor data, forcing the nose down while pilots fought to regain control. Less than five months later, Ethiopian Airlines Flight 302 crashed under nearly identical circumstances. 157 killed. Same failure mode. Same single point of failure. Same risk that had been flagged internally and classified as acceptable.

The fleet was grounded for 20 months. Boeing's total costs exceeded $20 billion. Congressional investigations exposed a certification process in which Boeing employees effectively supervised their own safety assessments on behalf of the FAA.

The stated metric was zero accidents per million flight hours. The actual metric — the one that drove daily decisions, resource allocation, and bonuses — was days to FAA certification. One measures safety. The other measures speed. When those two metrics diverge, the one attached to quarterly targets wins. Every time. 346 people paid for that divergence.

"The dashboards were green. The mission statement said safety. People died anyway. That is what execution misalignment looks like at scale."