What You’ll Find This Week

In December 2008, a few months before Uber was founded, TechCrunch reviewed a new iPhone app that offered "on-demand cab service from your iPhone at the push of a button." It was called Taxi Magic, it already worked in more than 25 US cities, San Francisco included, and it was built for taxi companies.

The taxi industry had the app first. It lost anyway, and Uber spent $32.9 billion making sure of it.

This week: why Clayton Christensen's own theory said the taxis should have won, the one thing Uber had that they couldn't match, and who has been paying the bill since.

Don’t Miss Our Latest Podcast

This Week’s Article

In December 2015, Clayton Christensen, Michael Raynor, and Rory McDonald used the pages of Harvard Business Review to settle an argument about Uber.

"Uber is clearly transforming the taxi business in the United States. But is it disrupting the taxi business? According to the theory, the answer is no." (What Is Disruptive Innovation?)

Disruption, in Christensen's theory, starts at the bottom of a market or with people who weren't buying at all. Uber started in neither place. It launched in San Francisco, which the authors called "a well-served taxi market," to customers "already in the habit of hiring rides," and the ride was better than a cab from day one. That makes Uber a sustaining innovation, and incumbents facing a sustaining innovation usually fight back and win. The authors even listed how the taxi companies were fighting: "They are deploying competitive technologies, such as hailing apps, and contesting the legality of some of Uber's services."

Christensen made the same call about the iPhone in 2007, when he said Apple had launched "an innovation that the existing players in the industry are heavily motivated to beat." The phone makers lost that fight.

The taxis were motivated too. They fought Uber with apps of their own and by challenging whether Uber was legal. And by Christensen's rules, they should have won.

In December 2008, TechCrunch reviewed an app that, in its words, was "doing what the others can't: on-demand cab service from your iPhone at the push of a button." A rider opened the app, it found them by GPS, and it sent the request straight into a cab company's own dispatch system, with no call to an operator. It worked in more than 25 US cities, including San Francisco.

I know what you’re thinking, but Uber wasn’t founded until 2009 and didn’t offer their first ride until 2010. Taxi Magic, the app reviewed by TechCrunch, had been running in San Francisco cabs for about a year and a half before Uber ever gave its first ride, and it was built for the taxi industry.

That’s right, the taxi industry had a working ride-hailing app before Uber existed.

Taxi Magic rebranded as Curb in 2014 and reached 60 cities. In San Francisco, a second taxi app, Flywheel, was running on about 80% of the city's cabs by the end of 2015. The taxi companies got there first, but it didn't slow Uber down.

Uber's advantage came from somewhere other than the phone.

UberCab launched in San Francisco as a way to summon a black car driven by a licensed limo driver. In July 2010, TechCrunch reported that a ride could run "one and a half to two times the price of a cab fare." That October, San Francisco's transit agency and the California Public Utilities Commission ordered UberCab to stop. One of the taxi industry's complaints was that UberCab offered immediate pickups, which the law reserved for licensed taxis.

In July 2012, Uber moved down-market. UberX launched in San Francisco and New York with 50 to 100 hybrid cars in each city, at a lowest estimated average fare of $14, against $24 for a black car. UberX's first drivers were existing Uber drivers, who agreed to buy the hybrids themselves. Travis Kalanick told AllThingsD he expected a good amount of cannibalization of the black-car business, and he described the trade plainly: "The best way to describe it is that the experience will be efficient, but not as elegant."

Uber launched the expensive version first and the cheaper one second. Researchers Glen Schmidt and Cheryl Druehl named that sequence high-end encroachment. It's how the iPhone took the phone market from Nokia and BlackBerry, and it's the order Tesla followed on purpose. Christensen described Uber's version himself, in the same article that ruled it out as disruption: "Uber has gone in exactly the opposite direction: building a position in the mainstream market first and subsequently appealing to historically overlooked segments." He saw the direction and his theory had no name for it.

The supply of drivers moved down-market next. Lyft and Sidecar were already letting ordinary people drive passengers in their own cars, and the utilities commission's enforcement division had issued cease-and-desist letters and $20,000 citations to Uber, Lyft, and Sidecar for operating without authority. In 2013, it signed settlements with all three services that let them keep running while the commission wrote the rules. In February 2013, Uber opened UberX in San Francisco to unlicensed drivers, screened by Uber with a background check and an in-person interview. A black car still needed a licensed limo driver. UberX now needed a person with a car who passed Uber's screening, and California had no license category for that person.

On September 23, 2013, the commission created a new category, the "transportation network company," defined as a service that uses "an online-enabled application (app) or platform to connect passengers with drivers using their personal vehicles." The decision named UberX. The new rules required background checks, driver training, vehicle inspections, and $1 million in liability coverage per incident. None of them limited how many drivers could sign up.

A taxi medallion is a license to operate a cab, and cities cap how many exist. A fixed supply is what let medallions sell for seven figures. In March 2014, at what turned out to be New York's last medallion auction, the high bid worked out to $1,240,250 per medallion.

New York's Taxi Medallion Task Force tracked what happened next in its January 2020 report. Medallion taxis went from 485,000 trips a day in 2014 to 296,000 by early 2018, while app companies averaged 600,000. By March 2018, the city had about 130,000 licensed for-hire vehicles and was adding roughly 2,000 a month. And by November 2019, the average medallion sold for just $164,518. Many owners had borrowed to buy their medallions, and the loans didn't shrink when the price did. In a city survey, the median owner-driver still owed about $499,000 on a medallion that now sold for about a third of that.

The task force put part of the blame on the medallion market itself. Speculation, risky lending, and weak regulatory oversight had led owners to take on heavy debt. When the app companies cut into taxi trips, medallion values fell and that debt became more than many owners could carry. More than 950 medallion owners had filed for bankruptcy.

The part Christensen's framework missed is why the taxi companies couldn't respond. Matching Uber's supply would have meant asking for more licenses, which would have cut the value of the scarce asset most taxi owners had borrowed against. The taxi companies already had the app, and they could challenge whether Uber's services were legal. But they couldn't copy Uber’s supply advantage without writing down the medallion.

Christensen's own article explains that kind of paralysis, noting that full substitution can take decades "because the incremental profit from staying with the old model for one more year trumps proposals to write off the assets in one stroke."

In March 2022, Uber signed a deal with Creative Mobile Technologies, which runs the technology inside New York's yellow cabs, to list more than 13,000 city taxis on the Uber app. And that November, hundreds of Flywheel taxi drivers in San Francisco started accepting Uber rides through a screen that switches between the two apps. The city had more than 700 medallions, and 350 had been foreclosed in the previous five years.

"Because the business is dead you know. We don't have a business like we used to," driver Wahab Alhindawi told ABC7. Flywheel's general manager, Muwaffaq Mustafa, put the deal more diplomatically: "It's better for us to work with them and come together and serve the city."

Flywheel was the software that 80% of San Francisco's cabs ran on in 2015. Seven years later, it was routing those cabs into Uber.

Uber didn't get there cheaply. Its accumulated deficit, the running total of everything it lost since its 2009 founding, peaked at $32.9 billion in March 2023. Uber didn't post a full year of operating profit until 2023. For more than a decade, investors subsidized the gap between what Uber charged its customers and what the business cost to run.

As Uber's position firmed up, fares inevitably climbed. According to Second Measure and YipitData figures compiled by Len Sherman, Uber's average US fare per trip rose 83% between the start of 2018 and the third quarter of 2022, about 17.5% a year while consumer prices rose about 4.5% a year. In 2025, ride-hailing prices rose another 9.6%, according to Gridwise Analytics, which tracks driver trip data. Comparing December 2025 with December 2024, the share of each fare that Uber and Lyft kept rose 33.2% per trip, while driver pay per trip rose 3.6%.

Consumer Reports found in March and April 2026 that Uber and Lyft keep between 43% and 49.5% of what passengers pay, and that riders requesting the same trip at nearly the same time saw a median price spread of 42.4%. Uber disputes the second finding, saying a trip "is defined not only by where it starts and ends, but also by when it is requested."

Uber raised more than $24 billion from private investors before it went public in 2019, then another $8.1 billion in its IPO. Its operating profit has roughly doubled every year since it turned positive: $1.1 billion in 2023, $2.8 billion in 2024, and $5.6 billion in 2025. As of June 30, 2026, the accumulated deficit was down to $7.95 billion. On paper, Uber has made back roughly three-quarters of what it lost.

But a large share of that recovery came from accounting rather than operations. About $11.4 billion of it came from tax benefits in 2024 and 2025, when Uber recognized that its old losses could now offset future taxes and booked that as income. That's standard accounting, and nothing about it is improper, but no rider paid for it. Operating profit from 2023 through 2025 adds up to about $9.5 billion, less than a third of the $32.8 billion deficit Uber carried into 2023. Uber spent $6.5 billion buying back its own stock in 2025 and another $3.5 billion in the first half of 2026 anyway. (These figures cover the whole company, including Uber Eats and Freight, not just rides.)

At 2025's pace, with no further growth, operating profit would close the remaining gap around the end of the decade. (That's a rough projection, not one Uber has made.) If the projection holds, Uber's strategy will have paid for itself about twenty years after the company was founded.

Travis Kalanick didn't wait twenty years. He resigned as CEO in 2017, sold about $1.4 billion of his stock to SoftBank in January 2018, and sold about $2.5 billion more in six weeks at the end of 2019, more than 90% of what he had left. That's roughly $3.9 billion, collected four years before Uber's first profitable year. Co-founder Garrett Camp took a different path. By that December, he'd sold only about $50 million of his shares.

The people who carried the cost didn't get to sell early. Medallion owners who borrowed to buy their licenses watched them lose most of their value, and in 2019 the median owner-driver in New York still owed about $499,000 on a medallion that sold for about a third of that. Riders paid 83% more per trip in 2022 than at the start of 2018, and fares kept climbing. Drivers saw their pay per trip rise 3.6% in 2025, the same year the share of each fare that Uber and Lyft kept rose 33% per trip.

Christensen asked whether Uber was disruptive and answered no.

But the question that mattered more was who would carry the cost that enabled Uber to outlast everyone else, and who would get paid before the bill came due. More than $32 billion in investor money made Uber possible, and it made Uber's founder a billionaire several times over, years before the company started earning back its losses.

Whether that's a model worth wanting more of is a question for another week.