Eighteen steps, each one a claim, in the order a partner would hear them

IndiGo carries roughly two thirds of every domestic passenger in India. That position generates the cash and the feed network any long-haul expansion has to be built on.

Domestic is not the problem. It is the platform.

benchmarking.carrier_operating_summary(), which returns the DOMESTIC table unless told otherwise. That default caused a published error once: an international exhibit plotted a 943 km domestic stage length. Here domestic is what is wanted, and it is named rather than defaulted.

Half of India's international traffic touches a Gulf point. The corridor is four times the size of India's entire direct market to Europe, and larger than Europe, North America, East Asia, Africa and Oceania combined.

Any conversation that starts with London and New York has already missed where the volume is.

Country-pair passengers bucketed to corridors through is_gulf_point(), never by membership of a literal list. DGCA writes ABUDHABI and RAS AL-KHAIMAH, and exact matching against a tidier list once filed 5.0M passengers a year, a fifth of the Gulf flow, under “everywhere else”.

Nothing about this number is in dispute, and that is the point: the scale is real. What it cannot tell you is where those passengers were going, which is the next exhibit and the one the recommendation actually turns on.

India's own statistics record the first foreign point and nothing beyond it. A passenger flying Delhi to Dubai to London is counted as a passenger to the United Arab Emirates.

That blind spot is not a data problem to apologise for. It is the finding. What the Gulf hubs do with those passengers after they land is where the margin goes, and India cannot see it.

Kochi sends more passengers to Gulf hubs than to the entire rest of the world combined.

Pivot 3. This chart was wrong for weeks. DGCA writes ABUDHABI where the code looked for ABU DHABI, so 5.0M passengers a year, a fifth of the entire Gulf hub flow, sat in the "everywhere else" bucket. All 72 tests passed the whole time, because a wrong bucket is still a valid bucket. How this analysis changed

benchmarking.gateway_flows(), splitting each gateway’s traffic into Gulf points and everything else. No modelling: it is a bucketing of the country-pair table, and the flows sum to the corridor totals in the exhibit above.

The diagram shows where the record stops, not where the passenger goes. How many of those Gulf-hub passengers continue elsewhere is the one number this project cannot measure directly, because no source publishes a Gulf six O-D table. It is gated UNVERIFIED_NO_PRIMARY and read in exactly two places.

India's own statistics put the Gulf at 51 per cent of international sectors. Origin-destination figures, which follow the passenger to where they were actually going, put it near 40. The eleven point difference is roughly eight and a half million people a year whose journey begins in India, ends in Europe or North America, and is sold by a Gulf carrier.

Priced across that journey at the only two yields this project has verified, its own and Emirates', the contested revenue runs between 28,900 and 56,700 crore rupees. A third to two thirds of IndiGo's entire annual revenue, in a pool it does not currently compete for.

Read it as the size of the contested pool, not a prize anyone captures whole. And note the weakest link is named: the 40 per cent has no Gulf six source. IATA's free report puts a wider Middle East at 39.2 per cent, which corroborates it and bounds the gap from below. It stays gated and everything drawn from it is reported as a band.

options.value_at_stake(). The 10.9 point gap between DGCA’s 50.9% sector share and a roughly 40% origin-destination share, times total traffic, priced across a 6,731 km reference journey. Reported as a band because the floor and ceiling are two different carriers’ realised yields.

This is the likeliest reason the case is wrong. The whole figure rests on gulf_od_share_pct, which carries UNVERIFIED_NO_PRIMARY and still cannot clear, because IATA's free report covers a wider Middle East and not the Gulf six. It corroborates the share at 39.2% and bounds the gap from below. If the true connecting share is materially smaller than eleven points, the contested pool shrinks with it.

Indian carriers account for fewer than half of India's international passengers. Gulf carriers take a quarter.

This is a market where the home team is still the minority shareholder.

benchmarking.who_carries_india(). Carrier names are matched through a normaliser rather than exactly: DGCA’s own file contains QATAR AIRWATYS, and a GRAND TOTAL row worth 17.53M passengers was once counted as a foreign airline.

The obvious version of this case is that India is losing its own market and needs rescuing. The data says the opposite. Indian carriers have gone from 37 per cent of international traffic in 2015 to 46 per cent in 2025, while the Gulf carriers have fallen from 33 per cent to 26.

That matters, because a recommendation built on a false premise is worth nothing. The gap is closing on its own. What has not closed is where it is closing: the share taken back is short-haul, flyable with the narrow-bodies India already has. The long-haul deficit needs an aircraft that has only just been ordered.

Pivot 2. This case opened on the opposite premise, that Gulf carriers were eating India's market and the wide-body order was a rescue. The trend reversed it. The finding only surfaced after a GRAND TOTAL row worth 17.53M passengers was found being counted as a foreign airline. How this analysis changed

Two years are dropped and the reason is stated rather than the gap left unexplained: repatriation and air-bubble flying put Indian carriers above 50% in 2020 and 2021 on a fifth of the usual traffic, which is a lockdown artefact and not a trend.

If the trend is read as “the problem solves itself”, the case collapses into doing nothing. It does not: the share taken back is short-haul, flyable with aircraft India already has. Watch the split by corridor, not the headline.

In 2025 IndiGo carried more international passengers than Air India while flying barely half the distance per passenger. Its international network averages under 2,700 km, which is the Gulf and Southeast Asia. Air India's averages over 5,300 km.

Neither airline is doing anything wrong. They are simply in different businesses, and the wide-body order is IndiGo buying its way into the second one.

International table, named explicitly. carrier_operating_summary() returns domestic by default, and reading that default here would show 943 km instead of 2,643 km and reverse what the exhibit claims.

The Gulf is 52 per cent of India's international passengers. Weight the same traffic by distance flown, and its share of revenue falls to 31 per cent. A Dubai sector is about 2,200 km; a New York sector is about 11,700 km, and revenue scales with the distance.

No margin assumption enters this chart. It is passenger counts, great circle distances and one published yield, so it stands even if you reject the profit model that follows.

No margin assumption enters this exhibit at all. It is passengers times distance times IndiGo’s published yield per RPK. It exists so the point survives a reader who rejects the modelled margin in the exhibit that follows.

Width is revenue, height is margin, so each block's area is that corridor's profit. The Gulf stays the widest block because volume is real, but the long-haul corridors sit above it, and that height is what a wide-body buys access to.

The margin axis is modelled, and labelled as such, because no Indian carrier publishes a margin split by corridor. It is anchored so the revenue-weighted mean equals IndiGo's 27.3 per cent FY2026 EBITDAR margin excluding forex. IndiGo reported 17.8 per cent including it. Both are stated, because the difference is a treasury outcome on dollar lease liabilities, not a route one.

Margin is spread across stage length by MARGIN_STAGE_SENSITIVITY, a module constant with a sensitivity() beside it, and anchored so the revenue-weighted mean equals the ex-forex figure. Both FY2026 margins are stated: 27.3% excluding forex, 17.8% as reported. Publishing either alone is the error this project has already made once, in the opposite direction.

The margin axis is the most heavily modelled thing in the repo, because no Indian carrier publishes margin by corridor. Reject it and the exhibit before this one carries the same point with no margin assumption at all. That redundancy is deliberate.

Unit cost falls as sectors lengthen, because the cost of a departure is spread over more seat kilometres. Scale IndiGo's published unit cost across the corridors and ask a different question: how far could fares fall before each stops covering what it costs to fly?

The Gulf has the least room of any corridor, and on these numbers it does not cover its cost today. Europe could absorb a fifth off the fare and still clear.

This chart is one of the three findings that moved the recommendation off "Gulf first". It does not say the Gulf stops mattering: the corridor is still where the traffic and the contested revenue are. It says the aircraft should not be pointed at it, because the sectors are too short to carry their own cost and there is no entitlement to fly more of them anyway.

This corroborates the chart before it from an unrelated direction. The profit pool models margin upward from an EBITDAR anchor; this scales cost downward from a published CASK. They share nothing but the corridor distances, and they rank the corridors the same way. A test asserts they keep agreeing.

Pivot 1. This chart, the bilateral ceiling and the capacity arithmetic together moved the recommendation. It was "reclaim the Gulf corridor first" until three separate lines of evidence said the aircraft cannot go to Gulf airports: no entitlement to fly them on, the worst unit economics on the map, and an order book sized for a network a quarter longer than this one. How this analysis changed

Headroom, not breakeven against a flat yield. Yield per RPK falls with stage length, so holding it constant across an 11,766 km sector flatters long-haul. Reporting the tolerance instead leaves the unknown on the reader’s side of the line. There is no net present value anywhere in this project for the same reason.

The result moves with the load factor chosen. At the international 81.1% the Gulf sits about four points under cost; at IndiGo’s system 84.8% it comes out near breakeven. The ordering across corridors does not change either way, and the ordering is what the recommendation uses.

Every wide-body carries freight under the floor, and this case has ignored it until now. Europe moves 65 kilogrammes of it per passenger against the Gulf's 17, so the corridor being recommended carries nearly four times the cargo of the one being stepped back from.

Resist the obvious conclusion. This is not a long-haul effect: the correlation between sector length and freight per passenger is minus 0.10, which is nothing. North America is the longest corridor on the map and sits near the bottom. Belly freight tracks what two economies trade, not how far apart they are, and a chart ordered by distance shows that better than one ordered by freight.

East Asia is the number to look at twice: 130 kilogrammes per passenger on a corridor carrying two million people, the thinnest passenger market here with the densest freight. Nothing else in this case examines it. No revenue is claimed anywhere on this chart, because no Indian carrier publishes a freight yield.

Physical units and no revenue leg, because no Indian carrier publishes a freight yield. The correlation between sector length and freight per passenger is -0.10, which is nothing, so this is presented as a caveat against an easy argument rather than as support for one.

Three methods, extrapolating India's own traffic history, fitting air travel against income across peer countries, and counting the seats the announced order books can actually fly, put the 2030 international market between 96 and 109 million passengers, up from 78 million.

The capacity method is the low leg, and it was withheld for most of this project's life because its inputs were not verified. They now are, from the manufacturers' own airport planning manuals and IndiGo's annual report. The band is reported, never averaged: the spread between the legs is the useful output.

Pivot 5. Note which direction verifying the gated numbers moved the answer. The new leg is the low one, so the band widened downward and the recommendation got harder to argue, not easier. A gate that only ever unlocks good news is not a gate. How this analysis changed

Reported as a band and never averaged: the spread is the useful output. The capacity leg was withheld for most of this project’s life because its seat and utilisation inputs had not cleared the gate. Unblocking it widened the band downward.

Three growth paths, anchored on rates India has actually recorded rather than chosen for symmetry. The bear case is the slowest sustained three-year stretch in the clean data; the bull case is capped at 12 per cent, because beyond that the constraint stops being demand and becomes how fast aircraft, crew and slots can be brought on.

The spread between them is 27 million passengers. Every path requires materially more long-haul capacity than exists today.

The bear case is the slowest sustained three-year stretch in the clean data, not a round number chosen for symmetry. The bull case is capped because beyond 12% the binding constraint becomes how fast aircraft, crew and slots arrive, which is a different model.

Count what the firm order books can actually fly and the answer inverts. The 140 wide-bodies on order carry roughly twice the capacity Indian carriers need to hold their current share of a 2030 market. On today's network shape they would fly half empty of purpose.

They only clear if the average sector lengthens by about a quarter, from 3,400 kilometres to nearer 4,300, or if Indian carriers take share well past parity, or some mix of the two. That is the recommendation restated in capacity terms: these aircraft are not bought to carry more of the same traffic, they are bought to carry it further.

Block speed and sector length here are computed from DGCA's own aircraft-kilometre and aircraft-hour columns, not assumed. Air India blocks at 698 km/h across a 5,316 km network and IndiGo at 656 km/h across 2,643 km, which is the physical relationship you would expect and a reason to trust the columns.

Capacity in ASK, not seats: a seat is not capacity until you say how far and how often it flies, and this case turns on how far. Block speed is computed from DGCA’s own aircraft-kilometre and aircraft-hour columns rather than assumed.

The book is treated as net additional capacity, which overstates it: some aircraft replace retirements already inside the baseline. No public retirement schedule exists for either carrier, so the overstatement cannot be sized, and the absorption requirement is therefore a floor rather than a ceiling.

No primary source says when the deliveries begin. The Airbus release that confirms IndiGo's sixty firm A350s states no schedule at all, so no start year is asserted anywhere here. Three plausible starts are run instead and the spread is the output.

The shape is the same in all three: a gap opens in the bridge years, widest between 2026 and 2028, and closes as aircraft arrive. Slipping the start moves the whole delivery curve to the right without changing where it ends up, which is why the phasing question is about what to fly in the meantime rather than about how much was ordered.

No start year is asserted anywhere. The Airbus release confirming the sixty firm A350s states no schedule, so three plausible starts are run and the spread is the output rather than a single line implying knowledge nobody has.

IndiGo's unit cost rose 4.66 to 5.00 rupees per available seat kilometre in FY2026, and the headline reads as an operating collapse. Bridge it and the picture inverts: fuel fell 0.18, genuine non-fuel inflation was 0.11, and currency added 0.41. The currency effect alone is larger than the entire net increase.

This is the same correction as the retracted margin claim in the methodology, made at the unit cost line instead of the margin line. A dollar-denominated lease book is not an operating problem, but it is a real one.

Pivot 4. This exhibit exists because an earlier version of this page claimed IndiGo's operating margin had halved, 22.3 to 14.0 per cent. It had not. The number came from a convention the company does not publish, the real margin improved, and the claim was withdrawn. The pressure is real but it sits in unit cost, which is what this chart measures. How this analysis changed

4.66 to 5.00, bridged: fuel -0.18, genuine non-fuel inflation +0.11, currency +0.41. Currency alone exceeds the net +0.34. Use CASK ex-fuel ex-forex of 3.00 rather than 3.52 whenever an FX lever runs, or the forex effect is counted twice.

This exhibit exists because an earlier version of this page claimed IndiGo’s operating margin had halved. It had not; the claim came from a convention the company does not publish and was withdrawn. The pressure is real, and it sits in unit cost.

Between 10 and 40 per cent of IndiGo's unit cost moves with the rupee, depending on how much fuel is bought at the dollar international price rather than the rupee domestic one. India publishes no split, so the exposure is reported as a band.

At the top of that band a rupee move costs more than the same move in fuel. And every line starts above breakeven, because FY2026 opened with unit cost at 5.00 against unit revenue of 4.99. Long-haul aircraft are bought into that, not out of it.

The exposure is reported as a floor and a ceiling because the share of fuel bought at the dollar international price rather than the rupee domestic one is not published. The interactive version of this exhibit is on the dashboard, where a control indexes a precomputed cube.

Every major carrier flies above 80 per cent full. The aircraft that exist are full. The question is what the next hundred fly, and where.

Read the direction before the level, because it cuts the other way. Measured against 2019, IndiGo is down 87.8 to 86.1 and SpiceJet 92.7 to 86.2; only Air India and Air India Express are higher than before the pandemic. An earlier version of this step said load factors had recovered past their pre-pandemic level across every major carrier. They have not, and the narrow claim carries the argument anyway.

Two points per carrier, four carriers, from app_export.load_factor_slope(), which uses the same window and the same four names as the static site’s version.

The conclusion the exhibit supports is narrow: the aircraft that exist are full, so demand is not the constraint. It does not support the stronger claim that every carrier is fuller than it was in 2019, which this data contradicts for IndiGo and SpiceJet.