As global ride-hailing startup Uber heads toward a possible IPO this year, Wall Street’s eyes will be on its financials. Revenues have continued to grow quickly for the eight-year-old Silicon Valley company, but the bottom line isn’t pretty: Uber was on track to lose about $3 billion in 2016 on net revenue of $5.5 billion, according to Bloomberg News. That’s remarkable for a startup that has raised more than $11 billion with scant capital costs — it does not own a global fleet of cars or much of other hard assets. Uber itself is valued at more than $60 billion.

Can Uber slow its rate of cash burn before losses start to threaten the company’s viability? On the surface, stemming the red ink doesn’t sound so hard. Since it does not own vehicles or employ drivers, the company saves a fortune in capital and workforce costs. But Wharton experts point to other substantial costs: In helping to create an innovative new market — the sharing economy — Uber spent a fortune training, recruiting and subsidizing drivers, giving away free rides so consumers would get hooked on the service, setting up a global system of local and regional offices as well as hiring lawyers to deal with lawsuits and regulators.

“I think Uber thought, ‘We have this platform — this app, this technology — that can be leveraged anywhere in the world, so let’s just go and conquer the world,’” says Wharton management professor Exequiel Hernandez, who wrote two case studies on Uber for his classes, based on interviews with executives. “What Uber underestimated were the costs that didn’t have to do with their technology and their business model, costs that have to do with the politics of being legitimate, [addressing] regulatory resistance and even cultural differences across markets.”

The idea for Uber came to co-founders Travis Kalanick and Garrett Camp one snowy evening in Paris when they could not get a cab. Camp’s first idea was to start a “limo timeshare service” where riders booked cars on-demand through a smartphone app to fill the down time of livery car services. Later, Uber expanded its options to add lower-price rides such as UberX, UberXL, UberSELECT; new premium services UberBLACK, UberSUV and UberLUX; and carpooling rides with UberPOOL. Uber now operates in more than 500 cities and 70 countries.

In helping to create an innovative new market — the sharing economy — Uber spent a fortune.

Uber’s innovation did not stop there. It applied the concept of “surge pricing” to its service — prices would go up when demand for cars in an area outstripped supply. The idea was that higher prices would prompt more drivers to come out until demand and supply reach equilibrium. Hernandez says Uber developed a dashboard that provided real-time supply-and-demand data in a city that helped managers see the high traffic areas they could send drivers to and adjust prices accordingly. Rides were booked by passengers using stored debit, credit or prepaid cards so they did not need to have cash on hand. Further, a bilateral ratings system rated both drivers and riders, creating a window into their behaviors.

While the practice of surge pricing isn’t new — airlines have been using it for years — it was new to car services. Uber’s brilliance was to get riders to accept the practice. “People got used to it” as long as the cost was transparent, open and fair, says Senthil Veeraraghavan, Wharton professor of operations, information and decisions. For example, riders ponied up $7 during busy times for a ride that normally costs $4. While there was an outcry initially when some New Year’s Eve revelers were hit with short rides costing hundreds of dollars, people have learned to plan for it. Uber also has come out with tips to avoid such surprises. “Uber tells people demand will be high so plan around it,” he says.

Wharton marketing professor Ron Berman notes that Uber operates in a classic “two-sided market” where Uber is the “market maker” between the driver and rider by providing a platform for them to connect. As such, he adds, there is a “strong network effect” taking place. “The more drivers there are on the road, the less time riders wait and the better service they get. [That brings in more riders, and] more drivers want to join Uber, since they know they will have high demand and less idle time.”

Bull and Bear Case

For this model to work, Uber needed to reach critical mass. The company’s pursuit of growth is largely the reason behind its rapid cash burn. Berman says that for companies to reach critical mass under these circumstances, they need to subsidize both sides of the market — pay them to join the system. He notes that the costs for Uber are high: “They are losing about $3 for each $1 they make.” However, getting to critical mass also typically results in a “winner-take-all effect,” which is what happened with Google in search and Facebook in social networks, he adds. “In this case, Uber’s strategy to try and grow as fast as possible at the expense of making a profit makes a lot of sense.”

Uber could be positioning itself to thrive in the long run. “Another way to see it is that Uber’s playing a very long game. We’re not used to startups playing such a long game. We’re used to startups eventually getting to something that makes them profitable, [going to] IPO and exiting,” Hernandez says. “The closest example we can find to Uber is Amazon, where [founder and CEO] Jeff Bezos was willing for decades [to prioritize growth over profits] — and even to this day, Amazon really hasn’t been a very profitable company.” Pundits also wrongly predicted Amazon’s demise.

Spending heavily to corner the market also makes sense from a regulatory standpoint. “If it were just about competing against [ride-hailing startup] Lyft and [Chinese rival] Didi, then there is some value to the argument that they’re burning cash too fast,” Hernandez says. “But the regulation angle actually justifies the cash burn.” Uber needs leverage, money and legitimacy in order to get regulators to accept its service even though it threatens the entrenched taxi industry in many cities. By subsidizing rides at first, Uber gets more people to use and like the service. That’s a major advantage when the company goes in front of regulators.

“It needs the public on its side and it needs ridesharing to be a significant portion of the economy so that regulators have an incentive not to kill it but to regulate it in a way that preserves jobs and infrastructure around ridesharing. That requires size,” Hernandez says. “You’re not going to the city council of New York and say, ‘We’re just a local New York company.’ Whereas if you say, ‘We’re everywhere and we have this brand and people love us,’ then you [become] the 800-pound gorilla.”

“We’re not used to startups playing such a long game.” –Exequiel Hernandez

According to Wharton management professor Tyler Wry, the cash burn is less problematic when considering Uber’s growth plans. “They’re trying to dominate that space. With that much capital on hand, they have the ability to expand rapidly, saturate different markets and effectively block out competitors without worrying about getting themselves into a liquidity crunch.” He adds that investors who agree with Uber’s spending strategy would be on board with the risks.

The bear case for the cash burn is that Uber’s tactic of lowering prices to get more riders, even if it means taking a loss, is problematic in the long run. In order to succeed, “that strategy would count on its dominance of the market after all significant incumbent taxi businesses exit,” says Arkadiy Sakhartov, a Wharton management professor. “I do not believe the strategy would be sustainable.” For example, he notes that it costs him $95 to take a taxi to the airport from his home, compared to $32 for Uber. The taxi industry knows how much it takes to make money on the ride. For Uber to make a profit on the same ride, it would have to lower its costs by three times, which is unlikely.

Uber’s rapid global expansion also doesn’t come cheap. “Uber is trying to take on too much. It’s burning too much cash in too many foreign markets,” Hernandez says. “It’s expensive to operate and compete in different markets.” In contrast, U.S. rival Lyft has pursued a U.S.-centric strategy and has chosen to partner abroad. However, Minyuan Zhao, Wharton management professor, says she would not compare the two because they are pursuing different goals, time horizons and attract different investors. While there are justifications to Uber’s strategy, a case could be made that the company might be “too ambitious for its own good.”

Another source of Uber’s cash burn are its investments outside of its core ride-hailing service for consumers. Uber has launched UberEATS, a food delivery service, UberCHOPPER for requesting a helicopter and UberFreight for long-haul trucking, among others. But these could be distractions. “Trying to execute on multiple fronts can dramatically increase the complexity of a firm’s operations and split its focus. This is mitigated to some degree when different business lines build on the same underlying competencies, but it’s still a concern,” Wry said.

Complexities of the Global Market

Hernandez predicted early on that Uber would not succeed in China. “I think the exit from China actually proved perhaps they were being too cavalier about where they went.” In August 2016, Uber left the China market after a bruising fight with local rival Didi, losing a reported $1 billion a year. Uber traded its Chinese operations for a 20% stake in Didi while the Chinese startup said it would invest $1 billion in Uber. One reason Uber failed is “they weren’t going to win the regulatory battle there. The Chinese government favors local companies when they compete against foreign ones in China,” he adds.

“With that much capital on hand, they have the ability to expand rapidly, saturate different markets and effectively block out competitors without worrying about getting themselves into a liquidity crunch.” –Tyler Wry

Uber also misread Continental Europe. “They didn’t understand that … in Germany, Spain or France, this idea that ‘we’re providing an opportunity for the driver to be an entrepreneur’ doesn’t fly because people care much more about labor security and labor protection,” Hernandez says. Meanwhile, Uber is doing well in Latin American markets such as Mexico despite the opposition from taxi drivers. “In Mexico, where taxis are old and dirty and … expensive, even the mayor of Mexico City was on Uber’s side,” Hernandez notes. “The value proposition Uber brings is different.” The Middle East also seems to be a robust market for Uber. However, Asian markets remain a mixed bag for now.

The larger lesson for Uber is that it needs to better understand cities before it decides to enter them. While Uber can run the tech platform from its headquarters, most facets of its operations are local: drivers, riders, regulators, prices, cultural practices, among others. Uber understood this dichotomy from the start, but its operating model veered between centralization on the tech side and hyper-localization on operations, Hernandez says.

Indeed, problems arose when the objectives of the team doing the centralized launch clashed with those of the local operations team. “The operations team … focused on creating a sustainable operation. In contrast, the performance of a launcher was measured by the speed with which they opened a new city,” he wrote in his Uber case study.

For example, when a launcher negotiated unsustainably high commission rates with local drivers to get more cars on the road quickly, an operations manager complained. (Uber typically took a 20% to 25% commission on rides, Hernandez says.) Later, Uber realized it could save money by reorganizing into regional hubs instead of being hyperlocal. Now, regional managers make decisions for cities.

At least, Uber’s famously pugnacious approach to regulations seems to be softening a bit. Hernandez points to CEO Kalanick’s 2015 conference speech in Germany as signaling a shift from the company’s aggressive “principled confrontation” approach. His speech was titled, “Uber and Europe: Partnering to Enable City Transformation.” More recently, Uber said it would share anonymous, aggregated trip data with city officials. Such travel data could help cities decide where to invest in infrastructure to alleviate traffic, among other benefits.

Uber’s Future

To be sure, Uber still faces many challenges. Additional regulations could curb its growth, and lawsuits, such as ones to classify its drivers as employees with benefits, could substantially increase its costs. It already lost one such case in London last fall and settled two others. Uber also faces taxation, such as an 18% value-added tax Russia levied on electronic goods and services provided by global tech companies that took effect this year. Uber has said it would reimburse drivers, but the additional paperwork prompted some to quit.

“By undercutting prices and heavily investing in its technology, Uber may accept losses now because it counts on the future advantage of combining its business format with driverless vehicles.” –Arkadiy Sakhartov

As for Uber raising its commission to offset higher costs, that could be problematic because it would come at the expense of the drivers, says Veeraraghavan. Drivers already bear the cost of fuel, car depreciation and insurance — and they could leave for a competitor if Uber takes more from them. Already, Uber has a problem with driver churn. Berman says half of its drivers become inactive after only 12 months.

Problems with drivers are likely what’s motivating Uber to invest in self-driving cars. However, that means Uber would have to own cars. “They’re replacing the labor cost with the capital cost,” Veeraraghavan says. “Is the capital cost cheaper? I’m not sure.” But Sakhartov sees autonomous vehicles as the most plausible fix to Uber’s financial model. “By undercutting prices and heavily investing in its technology, Uber may accept losses now because it counts on the future advantage of combining its business format with driverless vehicles,” he says. “Such vehicles are now predicted to dominate all other cars by year 2035. In that case, the cost structure of the taxi business will change substantially.” Without any driver costs, Uber would recoup its losses.

Despite its cash flow issues, Uber is by far the dominant global ride-hailing startup — and it is expected to stay that way. Hernandez believes that in many markets, there will be an oligopoly composed of Uber and perhaps one or two local startups — not more. “The barriers to entry [have risen.] You need a brand, you need cash,” especially since Uber, with its big war chest, is always in the background. Berman adds: “I don’t see the market having more than two or three ride-hailing apps that will be profitable in the long run, unless prices increase and go back to [being comparable with] taxi companies.”