Scott Kirby on United’s Fleet Bet, Starlink and Airline Economics

United Airlines CEO Scott Kirby says passengers are already changing flights to get Starlink. To him, the important distinction is not free Wi-Fi but fast Wi-Fi: an internet connection that actually works can influence which airline a customer chooses.

John Collison asks whether United’s large aircraft order during COVID was a dangerous bet. Kirby says he expected the pandemic to be deep and long, but did not accept the prevailing forecast that travel demand would remain permanently depressed.

As other global airlines retired wide-body planes, United ordered 200 wide-bodies and several hundred narrow-bodies, Kirby says. He told the board that competitors’ retreat created an unusual opportunity: United did not need a perfect forecast, only one that was not disastrously wrong.

Planes have long lives, so an order placed in a downturn will span several economic cycles. Kirby says United obtained favorable pricing because few others were buying, although the company still had to judge the risk of a prolonged recovery.

Collison asks why so many forecasts went wrong. Kirby points to two habits: extrapolating the present indefinitely and treating answers to surveys as reliable predictions of future behavior. Someone frightened during a crisis may accurately describe how they feel without knowing what they will do afterward.

Kirby extends the point to financial models that assume extreme outcomes are unusually rare. He argues that real-world tails are fatter than simple assumptions suggest, so unusual events deserve more attention when making large decisions.

Surveys remain useful for understanding how customers felt about something United has already done, Kirby says. He is much less willing to use them as instructions for what to build next. Collison adds that metrics should inform judgment rather than replace it.

Kirby gives the airline’s Starlink tracker as an example: instead of running a focus group, the team built something customers could try. AI and modern software can make that kind of experiment cheap enough to launch, observe and stop if it does not work.

Ben Thompson asks how much United’s ability to experiment depends on its technology stack. Kirby says the company used the pandemic to invest several hundred million dollars in modernization, moving away from mainframes and toward accessible, modern databases.

Kirby says a conventional return-on-investment calculation could not easily justify every part of that work. Its value was the ability to move quickly and build capabilities that would be difficult on an older stack. He dates the decision to 2020.

Although he has a computer science degree, Kirby says his more recent AI realization came from programming rather than summaries of calendars or documents. After three one-hour tutoring sessions, he tried building a project himself.

During a workout, Kirby used AI tools to build a sports-betting interface. He says it collected odds from several providers, let him choose a wager and sent it to the provider with the best odds. Being able to create the application himself changed his view of the technology.

The airline also had to manage the operational risk of changing a major legacy system. Kirby says United deliberately did not schedule several hundred flights during the transition window, so that a failure would disrupt fewer customers. The switch ultimately went smoothly.

Collison asks whether airlines have outsourced too much technology. Kirby says United has brought more systems in-house and owns its reservation system, a legacy of earlier decisions that left the airline responsible for supporting it.

Kirby’s frustration with seatback maps led him to another example. He wanted a map that stayed where he put it, rather than repeatedly switching to a globe. United is taking control of the map and broader entertainment software, he says.

Owning that software could allow the screen to remember a passenger’s preferences. Kirby also imagines celebrating a two-million-mile milestone in real time, giving both the traveler and flight attendants a memorable moment. He credits the book *Unreasonable Hospitality* with reinforcing the value of creating memories.

Thompson asks what happens when a pleasant surprise becomes an expectation. He describes missing a connection after assuming United’s connection-saving system would hold the next flight. A service that normally delights customers can disappoint them when they come to rely on it.

Kirby says the airline needs to communicate better when circumstances change. He expects seatback screens and the app to share information, so passengers can understand their connection and any change rather than receiving disconnected messages.

For weather disruptions, Kirby wants warnings to include useful options. Instead of merely announcing possible bad weather, United could show alternative flights and estimated probabilities of disruption, allowing the customer to make a more informed choice.

Kirby says he asked officers to make their job “wow”: create something a frequent flyer has not seen before. The company gets credit for the surprise once, he says, and then has to keep improving because customers understandably take good features for granted.

The hosts discuss airline rankings that emphasize meals and service. Kirby praises Asian and Middle Eastern carriers on those dimensions but argues that technology can be a differentiator, particularly when a flight is delayed or a connection becomes uncertain.

In one weather-delay example, Kirby says United sent some passengers a radar image showing a line of storms along the flight path, while others got a conventional notification. He reports a large difference in customer scores and argues that people respond better when they understand the reason for a delay.

Collison calls this expectation management. Kirby prefers to call it communication: he does not want a known problem to become an excuse for accepting a worse experience. The airline should explain constraints while still trying to reduce their impact.

Kirby describes simulations for summer thunderstorms around Denver. He argues that better forecasts and routing might keep parts of the airport operating when weather affects only some approaches. His proposed reduction in delays is an estimate, and changes would require cooperation with the FAA and safety rules.

The goal, Kirby says, is not to take risks with weather. It is to find safe ways to reduce disruption instead of simply treating bad weather as an explanation that ends the discussion.

Asked about the main limit on United’s business, Kirby answers air traffic control. He believes it is the dominant constraint and that modernization is possible, but says the technical project is intertwined with politics and government procedures.

Kirby argues that annual appropriations are poorly suited to infrastructure projects that take five to ten years. Funding delays and the need to wind a project down at the end of a budget period, he says, interrupt work that ought to be continuous.

The hosts ask whether airlines could help fund improvements. Kirby describes the politics of user fees and says private aviation interests resist changes. These are his assessments of the funding system and its political obstacles, not an independent accounting of who pays what.

Kirby argues that the distributed organization of air traffic control creates inefficiencies, including vulnerability when equipment fails. He says he is not advocating FAA privatization because of its political toxicity, while maintaining that private-sector execution could move faster.

When air traffic control reaches its limit, Kirby expects the result to be fewer flights at more airports. He says restrictions have already affected New York and Chicago, even though the sky itself has room for more aircraft if the system can safely manage them.

Thompson asks why Kirby has argued that a larger United presence in Chicago could help customers. Kirby frames the answer in terms of a capacity-constrained hub and the number of flights needed to feed connecting passengers.

United’s short flights from places such as Milwaukee primarily feed the Chicago hub, Kirby says. When two airlines each operate similar feeder networks, both use scarce flight opportunities. He argues that a single larger hub could use bigger feeder aircraft and free capacity for more long-haul routes.

The hosts press Kirby on his rivalry with American Airlines and whether his former job there makes it personal. Kirby says the competition is about network economics, pointing to United’s share gains in Chicago and other markets. His characterization of rivals’ results is his own.

The conversation turns to Delta and Starlink. Kirby praises Delta’s work building customer loyalty, but says United is investing across the whole cabin. In his view, fast connectivity is becoming one of the most important reasons passengers choose an airline after schedule.

At the time of the conversation, Kirby says about 48% of United flights have Starlink and the share is approaching 50%. Installation began with smaller aircraft because they had received the necessary certification; other aircraft types require their own approvals.

Collison asks why other airlines have not adopted the same approach. Kirby says the service costs more but offers a much better experience. Initially, United planned to install it during heavy maintenance and cover only part of the fleet.

After seeing passengers choose flights specifically for Starlink, Kirby says United accelerated toward fleetwide installation. Thompson observes that a partial rollout does not provide certainty at booking: knowing the connection will be there can itself change a traveler’s plans.

Kirby says United plans to show customers an estimated probability of Starlink on a flight and aims for complete coverage in selected markets. He identifies Los Angeles as an early target, while noting that certification on Boeing 787s could slow some hubs.

Kirby thinks the industry’s traditional pace of aircraft modification made competitors underestimate the impact of a fast rollout. He expects the advantage to be temporary: eventually other airlines will also offer high-speed low-Earth-orbit connectivity.

The near-term opportunity, in Kirby’s view, is to persuade people to try United for connectivity and stay for the rest of its service and technology. He says the company has a few years to make that case before similar connectivity becomes widespread.

Asked about making Wi-Fi free, Kirby says United planned to do so anyway, but insists speed is the important benefit. A free service with insufficient bandwidth can be unusable when everyone connects. Thompson agrees that more users can make a limited connection worse.

Kirby acknowledges the service strengths of Asian and Middle Eastern airlines. He says United must keep improving beyond connectivity, and argues that being ahead can create resources for further investment rather than an excuse for complacency.

Collison asks how Kirby gets permission from investors and the board to make large bets. Kirby credits a supportive board and regular communication, while saying Wall Street often dislikes the company’s biggest investments, including aircraft orders.

Kirby argues that pressure to reduce unit costs can cause companies to cut investments whose benefits arrive later. His answer is to make the long-term decision and accept accountability: if a CEO makes a major bet that fails, he says, losing the job is a legitimate consequence.

With roughly 120,000 employees across very different roles, Thompson asks how Kirby manages the organization. Kirby says the CEO’s central job is building the right culture. He describes asking Los Angeles employees to win over travelers who try United because of Starlink.

Collison asks whether the airline industry’s reputation as a chronically bad business has changed. Kirby says the picture improved but remains uneven, and he forecasts losses for several U.S. airlines. He sees United and Delta as having built stronger loyalty than some competitors.

Kirby returns to a rule he considers fundamental: an airline cannot indefinitely fly routes that structurally lose money. He argues that pride can keep management committed to weak routes, even when the underlying economics suggest they should stop.

The hosts ask how United chooses new destinations such as Nuuk, Greenland. Kirby credits route planner Patrick Quayle and his team rather than his own intuition. They look for opportunities that historical airline data may miss because nobody has tried the route before.

Experimentation also means stopping when a route does not work. Kirby cites Bergen, Norway: United tried the destination, found it did not perform as expected and stopped flying it. Aircraft can be redeployed, limiting the cost of an unsuccessful experiment.

Kirby says the team explores demand by combining outside information with travel patterns. For destinations associated with pilgrimages in Spain and Portugal, it looked at travelers arriving in one city and departing another, then identified an opportunity to serve both ends.

Collison asks about aircraft progress: modern planes are much safer and more fuel-efficient than earlier generations, but not much faster. Kirby says he finds the pace of new airframe development frustrating and wants more competition among manufacturers.

Kirby discusses possible improvements including blended-wing aircraft and engine technology. United has invested in JetZero, which is working on a blended-wing design. He notes that a wider aircraft could also require changes to airport gates.

The conversation moves to onboard products, including a planned economy seating option that can form a flat surface. Kirby frames such features as part of building a brand and making travel easier for families, even if an individual feature is small in the company’s financial results.

A marketing example illustrates another cultural point. Kirby says a United executive acted quickly on a Starlink campaign without waiting for a chain of approvals. He argues that trust and a clear understanding of the company’s goals can give teams room to act.

Collison asks whether airport slot systems restrict competition. Kirby says scheduling more flights than an airport can handle is especially harmful to customers, and describes asking for limits at Newark. The hosts also raise the tradeoff between incumbent investment and access for new competitors.

Kirby argues that improved air traffic control could create more capacity and therefore more slots. He also says airport charges at major airports have risen sharply, making it harder for low-cost airlines to serve those markets.

The hosts ask whether large airlines themselves squeezed low-cost carriers. Kirby says United introduced basic economy and used larger aircraft so it could compete on price in many markets, while still offering its broader network.

Kirby’s explanation for Spirit’s failure centers on fares that did not cover costs. He contrasts this with Ryanair, which he says has retained the discipline of using lower-cost airports. His account is an industry participant’s interpretation, not a complete financial analysis of either carrier.

Thompson asks whether the pursuit of growth distracted low-cost airlines from their original model. Kirby reframes the lesson as knowing one’s strengths and staying disciplined. Large network airlines have also failed, he says, when they tried to imitate low-cost carriers too closely.

Ryanair’s discipline impresses the hosts: it declines opportunities that do not fit its model. Kirby applies the same principle to United, whose strength is its hub network rather than attempting to operate every possible point-to-point route.

A network competitor has an advantage on a route when it can fill the aircraft with connecting passengers. Kirby uses Dallas as an example: a carrier without American’s connections there would compete for a smaller local market.

Opening a new hub is a rare, difficult decision. Kirby says he has closed hubs during his career but has not opened one, and argues that building the scale to compete with established hubs can require airport capacity that is simply unavailable.

United also looks for incremental international connections. Kirby describes discovering demand for onward travel from Tokyo to Ulaanbaatar on a particular flight and adding service, alongside connections provided by its partner ANA.

Kirby tells aspiring airline CEOs that the route network is the foundation of the business. He is skeptical of keeping structurally unprofitable routes merely because someone calls them strategic. In his view, management must inspect route-level economics and be willing to stop flying.

Asked whether credit cards are the main source of airline profit, Kirby says that interpretation overlooks the cost of carrying passengers who redeem miles. He estimates that tens of thousands of people fly on United each day using miles; revenue alone is not the contribution after costs.

Kirby views frequent-flyer changes through the lens of brand loyalty. He says United invested in lounge space during the pandemic and must balance access with the quality of the experience, including limited first-class seats and crowded clubs.

Airline alliances extend a network beyond the places United serves itself. Kirby cites connections through Lufthansa’s Frankfurt and Munich hubs, but says the experience can feel inconsistent when a customer moves to a partner airline.

Kirby says he pushed partner airlines to make seat assignments available to United customers, eventually making it a condition of the relationship. His goal is a more consistent experience across the journey, even when a different airline operates part of it.

Asked about consolidation, Kirby says he wants to build an American airline that can compete with major international carriers. This is his strategic vision; the broader effects of consolidation on prices and competition are not resolved in the conversation.

The hosts ask how Kirby runs a global airline while raising seven children and reading widely. He says he limits meetings to roughly two or three hours a day, avoids spending time on social media and prioritizes sleep.

Kirby describes sleeping about eight and a half hours a night. He learned the importance of rest while counting cards: when he slept too little to spend more time at the tables, he found himself rechecking the count and worried he was making mistakes.

The episode closes with an anecdote about a Las Vegas casino that still would not allow him to play blackjack years after he was barred. Kirby and the hosts joke about how casinos changed the rules and odds, then thank one another for the conversation.