As a myriad of scandals can dominated coverage of Uber over the past several months, the privately held company released details of their financials perhaps to test investor sentiment or to change the dominant story about the on demand transportation service.

Here’s the most shocking thing — they’ve gone from losing a billion dollars a year, to two billion dollars a year, to about a billion dollars a quarter while gross bookings total now $20 billion a year.

The company lost an adjusted $2.8 billion in 2016. That figure rises to a $3.8 billion tally when losses related to Uber’s Chinese operation are included, according to a Bloomberg estimate. (Those results compare to loss of “at least $2 billion” in 2015, according to a previous Bloomberg report). Both figures do not “account for employee stock compensation, certain real-estate investments, automobile purchases and other expenses,” according to Bloomberg.

So, on an adjusted basis, Uber lost around $3.8 billion in 2016. The real figure, using a full-GAAP reckoning, is likely higher.

Using the $3.8 billion loss figure against $6.5 billion in net GAAP revenue they had a -58% margin.

They want to argue they’re growing so fast that losses don’t matter. Indeed, they report revenue up 74% in the fourth quarter — except that when they complete a traditional ride they book only their cut as revenue (which works out on average to nearly a third) while when they complete an UberPOOL ride they book the whole ride as revenue and driver payouts as expenses. So as UberPOOL grows, their revenue grows disproportionately.

Uber can keep lighting money on fire for awhile, they have $7 billion in cash and access to significant credit lines, plus the possibility of tapping further private investment or going public. Whether they can do so at a valuation higher than — or rather significantly higher than — their current $68 billion valuation remains an open question.

They’re spending a lot to subsidize drivers as they grow aggressively into markets. They keep prices low, but guarantee income. They want to onboard drivers so there’s always a car available when customers pull up the Uber app — otherwise customers in a new city will get frustrated, and not use them again. And they want to be the low cost provider in the market.

I once described Uber as being like Saturday Night Live’s First CityWide Bank of Change. ‘They lose money on every ride, but make it up on volume.’

A lot of people don’t realize that change is a two-way street. You can come in with sixteen quarters, eight dimes, and four nickels – we can give you a five-dollar bill. Or we can give you five singles. Or two singles, eight quarters, and ten dimes. You’d be amazed at the variety of the options you have.

…All the time, our customers ask us, “How do you make money doing this?” The answer is simple: Volume.

There are advantages to scale. You want a big customer base to market to. Lots of customers mean drivers want to be on your platform. And lots of drivers mean customers get served quickly so they’re unlikely to be frustrated and want to switch to another platform.

It’s not like being the big social media network though — Facebook — that’s valuable because all of your friends are already on it, and it’s a pain to switch because your photos are there too. And even Facebook wasn’t the first social network. No one uses Friendster or Myspace anymore.

There are hurdles involved starting a new ride hailing service. You need to convince drivers to get on the platform, but you need customers or else you need to guarantee them income for signing up. But you can’t get the customers unless you have a critical mass of drivers. You can overcome these challenges with a significant amount of money to fund the launch and enough pre-planning.

Indeed, while Fasten and other services didn’t perform well in Austin during South By, the fact that very ‘similar to Uber’ services work in Austin now that both Uber and Lyft have left suggest that the technology isn’t a barrier to entry.

And to make money Uber is going to need to raise prices and stop offering generous incentives, and that’s going to make it easier for new entrants to compete.

• There’s really very little effort for consumers and drivers to run more than one app. Customers just have to download the app.

• In Austin (where Uber and Lyft have left the market) drivers I talk to all have multiple platforms that they drive with, including several running at the same time.

• And Uber doesn’t dominate everywhere it goes, not only did they sell out to Didi in China they’re in Berlin and Munich but not Dusseldorf, Hamburg, Frankfurt, or smaller cities in Germany.

There’s a benefit to being first and to being the biggest. But that doesn’t guarantee future success. Indeed, Uber’s negative margins may be squeezed even further as local regulations develop and drive up costs — whether in-person fingerprint background checks or Massachusetts taxing every ride and literally giving the money to cab companies. (Why does this remind me of the auto industry bailout?)

The intersection of Taxis and The State, Arlington Virginia

To justify the valuations that Uber’s been getting it may need to do more than dominate ridesharing. It may need to win not just play in transportation logistics including delivery, and the future of self-driving cars. They’ve got those cars on the road but it’s not obvious that they’ll ultimately dominate Google, Apple, Tesla, and other entrants.