This is not a trick question. Nor is it a query submitted to a fantasy writer’s listserv.

It’s a serious issue that is occupying Silicon Valley investors. Privately held tech startups worth more than $1 billion are known in the trade as “unicorns.” There are a lot of them around these days — so many that people are starting to ask whether this isn’t something of a bubble.

The core issue is that many of these unicorns aren’t profitable. They’re huge and they have lots of customers. They are the subject of both glowing write-ups and angry screeds about the “gig economy ... But every sale they make costs them a bit of their backers’ money. It’s the high-tech version of the old economist’s joke: “We’re losing money on every unit, but we’ll make it up in volume!”

Of course, that strategy did work for at least one company: Amazon, which is still just barely bothering to turn a profit as it heads into its third decade. Investors are willing to let the company skate along the break-even line because for decades, it was using that money to build up its market position. At this point, it’s hard to imagine anyone putting together an operation that could compete with Amazon, because the investment would be colossal.

The company can site and build a warehouse in under a year, and its combination of technology and logistics generates substantial economies of scale that would make it very hard for an upstart to compete on price. Investors believe, with decent reason, that this means Amazon will be turning solid profits for many years to come.

Unicorns obviously hope to pull off a similar trick: acquire the customers now, and later, having secured a nice cozy monopoly, raise prices to monetize them. There are some reasons to be skeptical of this approach, however.

A lot of companies tried the Amazon strategy during the dot-com era; exactly one company managed to leverage it into tech-giant status. Many more ended up like Kozmo.com, which spent a lot of VC cash delivering stuff below cost, then abruptly went out of business when the money ran out.

Let’s think about the easiest ways to protect an acquired monopoly, or oligopoly:

• Barriers to entry. It costs a lot to enter the market (think auto manufacturers) or you have some sort of government-granted protection from competition (patents, copyrights, local telecoms or cable monopolies, professional licensing rules).
• Switching costs. How expensive is it for me to switch to your competitor? For restaurants, the answer is “zero”; I can try another restaurant, for about what it would have cost me to eat at yours, and if I like that place better, I can keep going there. The only cost is the risk of a less-than-tasty meal. On the other hand, if I want to switch cell phone providers, I need to pay AT&T a lot of money to break my contract. And if I wanted to switch from Betamax to VHS, I had to buy all new videotapes to go with my whizzy new player. The higher the switching costs, the more likely customers are to stay put.
• Network effects. It wouldn’t cost me much money to switch from Facebook to another social media site. But Facebook is where all my friends are. I don’t just want to sit around and “like” my own posts; I want to interact! That means that the network with more users will tend to keep them — at least until young people decide that your site has too many users to be cool, and your network starts catastrophically shrinking.
• Economies of scale. When industries are characterized by large economies of scale, it’s hard for upstarts to compete, because they can’t make the product as cheaply as the incumbents.

None of these strategies makes you invulnerable forever, but if you’re going to spend huge sums on customer acquisition in the hopes that you will (somehow!) be able to make money off of those customers later, you’ll want at least one of these factors on your side. Ideally more than one: For instance, network effects are most powerful when switching costs are high.

How many of the unicorns meet this description? When I look at a lot of them, I see at best weak network effects, unprotected by switching costs. That’s a hard place to be in, if you’re trying to monetize all those customers you just acquired.

Of course, companies have made this work. Google’s dominance was largely driven by network effects. But there’s an important difference: Google isn’t trying to get its customers to pay for its services. Had it done so, it might have abruptly lost all its expensively acquired market share.

Instead, the company got the users, and then sold their eyeballs to advertisers, who were happy enough to pay less money for better-targeted ads. It also invested in creating economies of scale in ad delivery, by allowing people to buy and target ads directly, rather than relying on human customer service or expensive middlemen.

Many of the new round of upstarts, on the other hand, are providing services, like transportation, that have a high marginal cost. You want someone driven around, you’ll need to pay the driver a decent hourly rate to drive them. That means they won’t be able to realize Google-like economies of scale. At the same time, the switching costs of trying a new service are pretty low. That suggests to me that at some point, they are going to have to raise prices substantially — and only then will we find out if people are willing to pay enough for it to make the company profitable.

This is why I always roll my eyes a bit when I see articles discussing how Uber, or some other company, is going to completely change the way we work/communicate/travel/insert activity here.  We don’t even know what a lot of these companies will look like when they are mature.

They may be able to sustain large demand at higher prices and revolutionize our lives (and the wallets of the folks who invested in them). Or they may settle into a profitable but not transformational niche with somewhat less usage but better margins. Or many of them may slither out of existence, taking their massive investments with them. Because it’s only in myth that unicorns are immortal.

This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.

To contact the author of this story: Megan McArdle at mmcardle3@bloomberg.net