The busiest airport in America is on an allowance.

For its busiest days this summer, Chicago O’Hare had more than 3,080 flights on the books, nearly 15 percent above last summer. In April the FAA read that plan and replaced it with a smaller number: 2,708 flights a day, arrivals and departures combined, from May 17 through October 24. On a peak day, that’s up to 372 flights that will simply never exist.

The airlines didn’t pick the number. United expected to cut about 200 flights a day at peak. American put its own figure at no more than 40. Transportation Secretary Sean Duffy’s explanation was blunt: “unrealistic schedules were set [at O’Hare] to dramatically exceed what they could handle.”

The arithmetic behind the cap was already public. Last summer, only 56 percent of departures left O’Hare without a delay. The FAA pointed at airfield construction: taxiways closed for rehabilitation, concrete work west of the terminal. And under the pavement problem runs a staffing one. The agency has nearly 11,000 certified controllers plus more than 4,000 trainees in the pipeline and is still working toward its targets, and controller shortages have been tied to flow restrictions and delays at busy facilities, O’Hare among them. The latest hiring window drew 6,000 applications in its first 12 hours, every one of them headed into the back of that long pipeline.

Then came the part nobody downstream planned for. In April, the FAA said it expected “significant progress” on construction and didn’t anticipate restrictions past the summer. Chicago’s aviation department publicly thanked the agency for “ensuring they do not extend beyond summer 2026.” In July, the FAA extended the cap through October 30, 2027.

Here’s the structure the order exposed. The schedule is the visible artifact everyone plans against. Tickets get sold against it, crews get rostered against it, connections get promised against it. And the schedule was never really the airlines’ to set. It was an allowance, sized to what the system underneath (runways available, airspace, certified people in the tower) can actually move. When the plan and the throughput disagreed, throughput won. The only real choice was whether the smaller number got written down on purpose in April or extracted in delays all summer.

Your calendar works the same way. Somewhere in your operation there’s a 3,080 and a 2,708: the work you’re willing to book, and the work your real constraint can carry. The constraint is usually specific, and it’s usually a person. The one plumber whose license the inspector will accept. The single lift bay. The estimator, who is you, pricing every job after dinner. That resource sets your true daily capacity whether or not it appears anywhere in your plan, and it’s as slow to replace as a certified controller: you can collect 6,000 applications in a morning and still not have one more person who can work the scope this summer.

Sell past that ceiling and the cap arrives anyway. It just arrives in the ugly form: slipped dates, callbacks, the customer who books elsewhere next time. A backlog is a flight cap you didn’t choose, imposed by your own operation, without a press release.

The FAA’s move is the one worth copying, and it’s unglamorous. Find the constraint’s real weekly number and let that number set the schedule, before the backlog sets it for you. O’Hare’s cap still allows more flights than the airport actually flew on its best day last summer; a right-sized ceiling costs less than it appears to.

The extension carries the second lesson. A cap installed for one summer is now scheduled for two, and the parties who accepted it as temporary are issuing careful statements. Constraints outlast their occasions. If you ration on purpose, put a date on the ration, and actually reread it when the date comes.

So the question the cap leaves for your shop: what single scarce role quietly sets the ceiling on everything your business can take on, whether or not it appears anywhere in your plan?

← More observations