Two years of notice, thirty-five days to collapse

Two years of notice. Thirty-five days to collapse.

Written on 18 September 2026, based on public reports at that time.

In the first ten days of December 2025, India’s largest airline cancelled about 4,500 flights. Over ten lakh passengers were affected. The Railways added 116 extra coaches across 37 trains to carry people who could not fly.

It was not weather. It was not a strike. It was a rule about pilot rest.

The regulator had notified new duty and rest norms for pilots in January 2024. Nearly two years of notice. They came in two phases, the second from 1 November 2025. Everyone in the industry knew the date.

The airline collapsed about thirty-five days later.

The regulator’s inquiry named the cause, and the phrase is worth remembering. Over-optimisation of operations. Rosters built to squeeze maximum use out of every pilot, every aircraft, every route, with what the inquiry called minimal recovery margins.

The airline did not fail because it was badly run. It failed because it was run too tightly. Every spare pilot, every gap in the roster had been engineered away, because spare capacity costs money and looks like waste. So when the rules changed by a known amount on a known date, there was nothing left to absorb it.

The bill was roughly 500 crore in refunds, a show cause notice to the chief executive and the chief operating officer, and regulators sitting inside the airline’s operations centre. Months later the fix arrived, and it was almost embarrassingly simple. A deliberate crew buffer. They bought back the slack they had optimised away.

Now look at your own company.

One person who understands the export documentation. One machine with no backup. One supplier for a critical input. No bench anywhere. Each of those is a cost you removed, and each is invisible on your P&L until the week you need it.

You cannot carry slack everywhere. That really is waste. The discipline is choosing where.

So name your three tightest points. For each, work out what one week of failure would cost you in money and in customers. Buy a buffer only where that number is bigger than the buffer.

Efficiency and fragility are the same number. You find out which one you bought only when something changes.