Obviously, taxis will never be able to substitute for the people-moving power of high-capacity subways and bus corridors. Anyone who argues otherwise is being disingenuous.

But even for high-subsidy (read: expensive for taxpayers) transit routes, the budget-soothing promise of Uber isn’t cut and dry. What if grabbing an Uber isn’t about to get cheaper—but more expensive?

Hubert Horan, a transportation consultant writing at the blog Naked Capitalism, looks at some of the revenue tables Uber has provided to investors and concludes that the company’s current intake makes even American streetcars look like a good investment.

“Uber passengers were paying only 41% of the actual cost of their trips,” Horan writes. “Uber was using these massive subsidies to undercut the fares and provide more capacity than the competitors who had to cover 100% of their costs out of passenger fares.” The company’s profit margin in the year ending September 2015, Horan concludes, was negative 143 percent. It lost $2 billion despite revenues of $1.4 billion.

A large part of that was due to the subsidy battle Uber waged against Didi Chuxing in China, which ended with a merger this summer. But in the U.S., too, the company has lost money— “roughly” $100 million in the second quarter of 2016, according to investors who spoke to Bloomberg. On a per-ride basis, Lyft likely loses even more.

We’ve known for a while that Uber is unprecedentedly unprofitable, its $60 billion-plus valuation notwithstanding. But as we begin to make policy decisions based on it and its competitors’ impact, we have to recognize that this state of affairs can’t last. It is not just the taxi cartel that makes conventional cab rides cost more than Uber rides. It’s the patience and optimism of Silicon Valley investors. Maybe Uber will continue its shift into shared rides, which (as a prior generation of transportation operators learned 150 years ago) are more profitable. Or robot cars will eliminate driver jobs, dropping the marginal cost of providing rides (though adding billions in capital expenditures). But in any case, whether it achieves its desired market share or not, the company will have to start raising prices.

Right now, as the transportation writer Jacob Anbinder pointed out, Uber has a worse farebox recovery ratio—the ratio of price to subsidy—than most of the subway systems it is competing with. Not for long. Investors will expect profits—which means Uber rides will have to cost more. We’ll look back on these couple years as a magical era during which a heavily capitalized startup let us hop around town for pennies on the dollar. When prices come back up, Americans will find that cuts to transit service and reluctance to invest in expansion have left us in bad shape.

The irony is that Uber and Lyft are cognizant that being everyone’s private driver, at current prices, is not a sustainable business model. Both companies have pushed hard into services that resemble public buses, sometimes with appointed pick-up or drop-off points and fixed fares that are competitive with transit fares. (For now, anyway.)

But for some reason, transit officials see the imitation not as flattery—not as a validation of what they have been doing for years and a hint of how they might do better—but as an excuse to hang it up.