Scott Kirby: Why United Is Betting on Starlink, Routes and Its Own Tech

United Airlines CEO Scott Kirby tells John Collison and Ben Thompson that Starlink is the biggest competitive advantage in airline history, and that route discipline, not perks, decides which airlines survive. On Cheeky Pint, Kirby explains United’s contrarian aircraft order during COVID, why surveys cannot predict the future, bringing the airline’s tech stack in-house, why air traffic control is 90% of the industry’s problem, how Spirit failed while Ryanair thrives, and the time habits behind running a 120,000-person company with seven kids.

TLDW

Kirby took over United in May 2020 and ordered 200 wide-bodies and several hundred narrow-bodies while every other global long-haul airline shrank, betting the experts and their surveys were wrong about travel demand. He started a several-hundred-million-dollar project to move off mainframes with no ROI case, and says that modern tech stack now lets United do things rivals cannot, from AI weather forecasts for travelers to taking over its seatback software from Panasonic. Starlink is on 48% of United’s flights, heading to 50% within a week, and Kirby sees a two-to-three-year window before rivals catch up. The rest of the conversation is a masterclass in airline economics: why the route network is the foundation, why “strategic” money-losing flights should be cut, why credit card revenue is overrated, and why low-cost carriers priced out of big airports cannot survive.

Thoughts

Kirby’s Starlink case rests on one sharp distinction: “Free was never the value proposition. Fast is the value proposition.” He says he has flown all three of his big competitors on their free legacy Wi-Fi and found it terrible, because there is not enough bandwidth to give it away. That reframes Elon Musk’s requirement that airlines offer Starlink free as almost irrelevant, since speed is what changes behavior. Kirby hears customers say they are switching to regional flights because those have Starlink. What makes the bet interesting is that he openly calls the advantage temporary. Everyone will eventually have low Earth orbit internet, so the real play is to use a few years of being the only airline with it at scale to get people to try United, then keep them with the technology underneath. The hosts push on the weak spot: at 48% of flights, a traveler cannot count on it, and certainty is what moves bookings. Kirby’s answers, a probability of Starlink in the app and Los Angeles as the first hub at 100% in the first half of next year, show he knows the window is only valuable if United closes the gap to certainty fast, which depends partly on Boeing finishing 787 certification work.

The most useful lesson for anyone who runs a business is Kirby’s advice to new airline CEOs: start with the route network, because if it is broken nothing else can save the airline. Look at the route P&L, and anywhere you hear the word “strategic” attached to a money-losing route, cancel it without even asking which hub or city. It is a rule that applies far beyond airlines, where “strategic” often means a loss nobody wants to own. His discipline extends to United’s own ambitions. He has closed three hubs in a 30-year career and never opened one, because United would need about 600 flights a day to compete in the Southeast and no airport there has the gates or runways. He also takes apart the popular idea that airlines are credit card businesses with planes attached. The card revenue is real, but 50,000 people a day fly free on miles, and nobody counts what carrying them costs. That is an important correction to a meme that treats one side of the ledger as the whole story.

His explanation of why Spirit failed comes down to one piece of arithmetic. The New York airports now charge $52 per passenger, and Spirit’s average fare before it liquidated was $58, before fuel, aircraft, staff or maintenance. Big airport authorities, he says, have raised costs by something like 70% since before COVID and priced low-cost carriers out. The deeper failure was strategic. Southwest once flew to Oakland and Burbank instead of San Francisco and LAX, and Ryanair still avoids Heathrow and Charles de Gaulle, but every US low-cost carrier eventually chased big airports like Chicago, Newark and San Francisco. Kirby is less convincing when he credits United only with competing fairly, though. Basic economy and bigger planes were designed to match low-cost fares profitably, which is a big part of why the low-cost model lost its price advantage. Both things can be true: the low-cost carriers lost discipline, and United built products to take their customers.

Kirby names air traffic control as “90% of the issue” limiting airlines, and his description of how the government funds it is damning. Projects that take five to ten years are run on annual appropriations: the budget arrives late, three months go to ramping up, four months to work and the last three to shutting down because there is no authorization past 12 months. He praises the current transportation secretary and FAA administrator, but says the political incentives are fixed: members of Congress ask how many jobs a project puts in their district, so there are about 450 facilities instead of a few centralized ones with backups. He also points at private jets, which he says fly essentially free while airlines pay nearly all the fees, as the most powerful lobby against change. With demand growing about 5% a year and caps already in place in New York and Chicago, the consequence is plain: fewer flights than travelers want, because the sky has room but the system does not.

Kirby’s most striking accusation is about the aircraft makers. Asked why airframes have barely changed since the 737 first flew in 1967 at the same speed, he says Boeing and Airbus have quietly, though not officially, agreed that neither will invest in new airframes, because their investors do not want them to. He says he has begged both for more than 20 years to build a 757 replacement. That is a strong claim from the industry’s biggest customer, and it matters because it means the efficiency gains he does credit, 50% less fuel and 30 to 40 times greater safety, have come incrementally rather than through new designs. His answer, “we need a third aircraft manufacturer,” is unlikely soon, which is why United invests in startups like JetZero and its blended wing design. It also echoes his air traffic control point: a lot of what holds back air travel is not physics but institutions with no reason to move faster.

Key Takeaways

During COVID, United ordered 200 wide-bodies and several hundred narrow-bodies while every other global long-haul airline permanently retired part of its wide-body fleet. Kirby told his board United did not have to be right, just not totally wrong.
Kirby says two human mistakes drove the bad forecasts: extrapolating the present into the future and trusting surveys. He uses surveys only to judge things United has already done.
In 2020 United also started a several-hundred-million-dollar project to leave mainframes for modern cloud systems, with no ROI case, and Kirby says it now lets United move faster than rivals.
United and Copa are the only airlines in the world that own their reservation system, and United is taking over its seatback software from Panasonic to remember preferences and mark milestones like a passenger’s two millionth mile.
After three one-hour sessions with a tutor, Kirby vibe coded a sports betting tool in about 2.5 hours that compared odds across four sportsbooks, including Kalshi, and placed the bet at the best one.
Sending delayed passengers a radar map instead of a standard weather notice moved satisfaction from about negative 60 NPS to positive 60.
Kirby says AI simulations could cut 50% to 75% of Denver’s summer weather delays, but that needs the FAA to change how it restricts traffic.
Air traffic control is the biggest limit on growth, funded through annual appropriations that cannot run multi-year projects well, with about 450 facilities kept for political reasons.
Starlink is on 48% of United flights, moving to 50% within a week, with about three aircraft a day being fitted. Los Angeles will be the first hub at 100%, in the first half of next year.
Kirby says six of 10 US airlines will lose money this year, American will probably lose about $1.5 billion in Chicago, and United remains solidly profitable despite oil costs up $8 billion.
United will fly to 53 transatlantic destinations next summer, up from 23 a decade ago, and only two new routes failed. Bergen, Norway was one, and United simply stopped flying it.
New York airports charge $52 per passenger, against Spirit’s average fare of $58 before it liquidated. Kirby says Ryanair is the one successful low-cost carrier left because it stays at secondary airports.
About 50,000 people a day fly United free on credit card miles, which Kirby says is the cost side ignored by the “credit card company with planes” meme.
Kirby limits himself to two or three hours of meetings a day, avoids social media and TV, sleeps 8.5 hours a night, and is still barred from playing blackjack at the Bellagio after years of counting cards.

Chapters

0:26 Buying Planes While the World Shrank

When COVID hit, experts said travel demand would be down 50% forever, and every global long-haul airline except United permanently retired wide-bodies. Kirby thought the premise was wrong and placed the biggest aircraft order in history, telling his board that with everyone else shrinking, United only had to be not totally wrong. The order also got better pricing and helped Boeing restart the market. He compares the approach to Samsung investing through memory downturns, and blames the bad forecasts on extrapolating the present, surveys of frightened travelers, and models that assume normal distributions.

6:35 Owning the Airline Tech Stack

United doubled down on technology in 2020, moving off mainframes onto modern databases and the cloud, and Kirby says rivals without that stack cannot match what United now ships. His own AI epiphany came within the last eight months, when he coded a sports betting aggregator after three tutoring sessions. United owns its reservation system and is taking over the seatback software, inspired by Unreasonable Hospitality, to create memorable moments. He describes radar maps that turned angry delayed passengers into happy ones, weather-delay probabilities for alternative flights by Christmas, and AI simulations that could keep Denver running through summer storms.

22:02 Aviation’s Biggest Bottleneck

Kirby says air traffic control is by far the biggest limit on the industry, even with the most supportive administration and FAA leadership of his career. Annual appropriations make multi-year projects absurdly inefficient, private jets fly almost free while airlines pay the fees, and politics keeps about 450 facilities spread across districts. He thinks private enterprise could modernize it in two to three years but does not advocate privatizing the FAA, calling it politically toxic. With capacity growing about 5% a year, he expects more flight caps like those in New York and Chicago.

26:59 The Hidden Economics of Airline Networks

Kirby argues Chicago would be better off with one dominant hub airline, because United and American each run duplicate short “catchment” flights, such as about eight a day each to Milwaukee, that feed connections rather than serve the city. Single-hub cities like Atlanta and Dallas use those slots for more long-haul flights and drive more economic activity. He denies his battle with American is personal, pointing to share gains against Southwest in Denver and against Southwest and Alaska in the Bay Area.

30:19 Starlink

Kirby calls Starlink, after schedule, the most important reason customers will choose an airline. United started with smaller jets because they were certified first, planned to equip only 65% of the fleet, then went all in after hearing passengers switch flights for it. He says Delta’s slower approach was a mistake born of the industry habit of eight-year modification programs, that free Wi-Fi on legacy systems is unusable, and that United has a few years as the only airline with it at scale.

38:28 What Wall Street Gets Wrong About Aviation

Kirby credits wide reading for seeing the tails other people miss, and says Wall Street hated the aircraft order and wants cost cuts that only come from not investing. His advice to other CEOs is to stop complaining about short-termism, make the big bet, and accept being fired if it fails. He says six of 10 US airlines will lose money this year, mostly from structural losses on routes driven by ego, while United and Delta have built brand loyalty.

43:43 It’s the Routes

Route planner Patrick Quayle and his team find destinations that are not in the data, like Nuuk in Greenland and a pair of Spanish and Portuguese pilgrimage cities spotted through broken itineraries. Kirby says Boeing and Airbus have quietly agreed not to invest in new airframes and that the industry needs a third manufacturer, and he recommends the Skunk Works book The Impossible Factory. He also covers blended wing bodies, the coach Relax Row, a marketing team that ran Starlink ads without asking him, and why slot controls at Newark helped customers.

55:44 Why Spirit Failed but Ryanair Works

Big airport costs have priced out low-cost carriers, and US low-cost airlines abandoned the secondary airports that made the model work, while Ryanair stayed disciplined. Kirby says discipline for United means flying only to its hubs, cancelling any “strategic” money-losing flights, and accepting that a new hub in the Southeast is impossible. He explains why credit card profits are overstated, how United doubled lounge space during COVID, and how it forced Star Alliance partners to offer seat assignments. He wants United to become a flag carrier Americans are proud of, as Emirates is in Dubai.

1:15:38 Kirby’s Unusual Time Management Hacks

Kirby eats with a serving spoon because it is faster, holds only two or three hours of meetings a day, skips social media and TV, and sleeps 8.5 hours a night. He learned the value of sleep counting cards in Vegas, where tired play cost him money. He is still barred from blackjack at the Bellagio, and says counting cards is now impossible because casinos changed the rules.

Notable Quotes

“We don’t have to be right. We just have to be not totally wrong.”
Scott Kirby, on United’s COVID aircraft order

“Surveys are the absolute worst way to figure out what’s going to happen in the future.”
Scott Kirby, on why travel experts got COVID wrong

“Air traffic control is 90% of the issue.”
Scott Kirby, on the biggest limit on airlines

“Free was never the value proposition. Fast is the value proposition.”
Scott Kirby, on Starlink

“We need a third aircraft manufacturer, to be blunt.”
Scott Kirby, on Boeing and Airbus

“You can’t charge 58 bucks when you’re paying the airport authority 52.”
Scott Kirby, on why Spirit failed

“If you hear the word strategic flying, you don’t need to know what hub it is, you don’t need to know what route it is, you should cancel it.”
Scott Kirby, on his advice to new airline CEOs

Watch the full conversation here.