But from inside the bubble, these companies look like winners already. And why shouldn’t techies see them as such, when they’re adding hundreds of new employees each month, announcing plans for giant new offices in Scottsdale or Denver or Salt Lake City, and toasting themselves at industry conferences designed to further promote their images of success. It’s easy to nod and applaud. They should be celebrated. They should be imitated!

Uber, the $51 billion king of the unicorns, is among the most celebrated and imitated. The company set a record recently, surpassing the highest valuation Facebook (FB) reached as a private company.

But last week we learned a key difference between Uber and Facebook: One was very profitable at that stage and the other is very much not. Five years in, Facebook had healthy net profits of $200 million on $777 million in revenue. (It has maintained those profits, reporting 31% operating margins last quarter.)

Uber is nowhere close to turning a profit, according to financial information leaked to Gawker: In 2013, Uber’s fourth year in business, Uber lost $56 million on $104 million in revenue. Those losses nearly tripled in half the amount of time, while revenue growth only doubled: In the first half of 2014, Uber’s fifth year, the company lost $160 million on $101 million in revenue.

Pair that with a Bloomberg report which said Uber lost $470 million on $415 million in revenue over an undisclosed time period, and we have a clearer picture of Uber’s financial health than we did before. We don’t have access to any other metrics that might make these losses look better, like how much Uber spends to acquire each customer, or how profitable it is in the markets it has operated in the longest. (Uber’s own professional response to the report: “Shock, horror, Uber makes a loss. This is hardly news and old news at that.”) But more than anything, the numbers are a reminder that Uber’s success is not guaranteed.

This is the rub with all the successful-looking startups propped up by venture funding. It’s easy to forget, as private company valuations shoot to dizzying heights, that venture capital is designed for high risks and high rewards. Uber, like its VC-backed peers, is not playing it safe. If you wanted that, go find a nice lifestyle business, or start an agency, according to the thinking in Silicon Valley. But Uber’s big ambitions don’t make its financials any less unsettling.

Consider that Microsoft (MSFT) raised a total of $1 million in venture backing for a 5% stake (a $20 million valuation) in 1981. That’s approximately $2.9 million on a $52.5 million valuation in today’s money. By the time Microsoft went public in 1986, it enjoyed pre-tax profit margins as high as 34%. In other words, Microsoft had been sustaining its own growth for years.

It’s not a perfect comparison. For starters, Microsoft was selling to a small, fast-growing market of computer owners — less than 10% of households in the 80’s — compared with almost two billion smartphone users today. But it provides some perspective when you hear stories of startup founders with unproven business models demanding “one-on-one deals” — a $100 million investment on a $1 billion valuation. Or when you see that they’re actually getting those deals: In the first half of this year, more than 100 startups have raised funding rounds of $100 million or more, according to CB Insights. Compare that to 126 such deals in all of 2013, and just 23 in the second half of 2013.

Defenders of this jump in mega-rounds will correctly point out that the $50 billion invested in private tech companies each year is tiny compared to the trillions deployed by mutual funds and public market investors. Further, because it’s all happening in the private sector, regular Joe and Jane investors aren’t risking their retirement funds like they did in the dotcom bubble. (They might not be so impressed by Uber if it goes public with this kind of balance sheet.)

As long as Uber’s financials are private, it can be viewed as a winner. The company’s business model has been praised as disruptive, pioneering, and inevitable. Uber’s model is so revolutionary that legions of imitators have copied it. Startups that deliver on-demand services through an app use the analogy “Uber for X” as shorthand for their strategies. And there is an Uber for everything: Uber for groceries, Uber for house cleaning, Uber for moving, and Uber for lawn mowing. There’s an Uber for doctors, massages, haircuts, flowers, reservations, booze, marijuana, dog walking, cat sitting, and pizza. There’s Doughbies, the Uber for chocolate chip cookies.

If Uber has to burn tens of millions per quarter to grow its business, what does that mean for the countless “Uber for X” startups that followed Uber’s model because it looks like a sure win? What does that mean for the $4.1 billion invested into on-demand startups last year?

We are already seeing the cracks. Homejoy, an “Uber for maids” company that raised $39.7 million in funding, shut down in July, blaming a class action lawsuit against the way it classifies its workers. But one could easily argue that regulatory attention on worker classification was just the final blow to Homejoy’s already-failing business. Uber is currently trying to argue its way out of a similar class-action lawsuit.

Last week, on-demand personal assistant startup called Zirtual “paused” its operations, blaming market circumstances and financial constraints. Likewise, Good Eggs, an on-demand grocery delivery company, dramatically scaled back its operations, laying off 140 employees. According to CEO Rob Spiro, the company’s single biggest mistake was growing too quickly.

This year Silicon Valley’s most prominent investors have issued warnings about out-of-control spending at startups, while maintaining that this is not a tech bubble. Uber investor Bill Gurley and Lyft investor Marc Andreessen have been the most vocal. But they’ve carefully avoided naming which companies are guilty of burning so much cash. Perhaps it’s because their warning hits too close to home.