IN THE REAL WORLD a flywheel is a mechanical contraption that stores rotational energy. In Silicon Valley it has come to mean something else: a perpetual-motion business that not only runs forever but is self-reinforcing. Thanks to powerful network effects, the theory goes, a digital platform becomes more attractive as it draws in more users, which makes it even more attractive and so on. The end state is a venture that has gathered enough energy to self-levitate and throw off tons of cash.

The payout on one of the most richly-funded bets of the past decade or so revolves around whether ride-sharing and delivery firms—which once were part of something known as the “sharing economy” but are better described as the “flywheel economy”—can actually ever live up to their heady promise. The outcome will matter to more than just venture capitalists who backed their growth. Whether these flywheels do gather unstoppable momentum is also of interest to regulators worried about technology’s propensity for winner-takes-all business models, not to mention paid-by-the-gig workers caught in its cogs.

Consider the results of Uber and DoorDash, the largest Western ride-sharing and delivery apps respectively. The distorting effects of the pandemic—which severely limited ride sharing, but gave a big boost to food delivery—warped their business models in recent months. Nonetheless optimists will have seen plenty to cheer them. On November 4^th Uber proclaimed it was finally profitable, albeit only the flattering metric of “adjusted EBITDA”. Strong third-quarter figures from DoorDash, which were released on November 9^th, could fuel an already heady rally in its shares (the firm also announced the acquisition of Wolt, a Finnish food-delivery company, for $8bn).

But look deeper and evidence is mounting that business flywheels are not defying the laws of capitalism. The money that went into building them recalls the railway mania among other past speculative investment crazes. The nine firms that have gone public so far—Uber, its American rival Lyft; Didi, a Chinese ride-sharing app; and six delivery firms, from DoorDash and Delivery Hero, which is based in Berlin, to China’s Meituan and India’s Zomato—collectively raised more than $100bn. In most cases, the capital was intended to jumpstart those network effects and make market dominance a self-fulfilling prophecy. Seemingly bottomless pits of investors’ cash went to subsidising rides and deliveries to juice demand. This reached absurd points: a pizzeria could make money by ordering its own food for a discounted price on DoorDash (which then paid back the regular amount). To justify such profligacy, interested parties pointed to the huge “total addressable market”, another popular term in Silicon Valley. Bill Gurley of Benchmark, an early investor in Uber, argued in in 2014 that the firm could vie for as much as $1.3trn in consumer spending if one saw it as an alternative to car ownership.

Measured against such visions, the flywheel economy has proven a disappointment. To be sure, the nine listed flywheel firms are still growing nicely—at 103% on average in their latest reporting period compared to the same period the previous year. This explains why they are collectively worth nearly $500bn. But self-levitating they are not. Nor are they profitable. Sales for the group amounted to $75bn over the past twelve months and the operating loss to nearly $11.5bn.

As the firms have discovered, their businesses are less perpetual motion machines than real-world flywheels that inevitably lose energy to friction, says Jonathan Knee of Columbia Business School and author of a book entitled “The Platform Delusion”. The network effect in fact has proved much weaker than expected. Many users switch between Uber and Lyft. Drivers also flit between them, or to delivery apps, depending on which model offers the best pay. This bargaining power from both sides means the system does not become self-reinforcing after all.

Technology, too, has turned out to be less beneficial than anticipated. Data collected by the firms help optimise their operations, but are not the decisive factor some had hoped for. Regulators keep pushing back. In London they have forced Uber to pay drivers minimum wages and pensions. In San Francisco they capped the fees DoorDash can charge restaurants for delivering their meals.

Uber’s tortuous path to stemming losses should temper investor optimism. It eked out a profit of $8m on revenues of $4.85bn. That excludes important expenses that are unlikely to disappear, such as stock-based compensation. The company has crawled out of its sea of red ink mostly by slashing costs, shedding technology assets such as its autonomous-car unit, charging higher prices and increasing its “take rate”, the share of the fares it keeps. As a result, an Uber is now no cheaper—and often more expensive—than conventional cabs, plenty of which can be hailed via apps these days. What is more, the company is now more of a delivery service than a ride-hailing app: Uber Eats generates more than half of sales. Add its other newer and still far smaller services, such as groceries and package delivery, and Uber looks much more like an old-economy logistics conglomerate than a metaverse-era tech company that merits a market capitalisation of more than $85bn.

DoorDash, whose revenue has grown more than four-fold since the last quarter of 2019, was expected to suffer as more people dine out, but revenue held up in the latest quarter at nearly $1.3bn, although net losses more than doubled to $101m compared with the same period a year ago. Regardless, investors have pushed it to a punchy $65bn valuation, up 101% since it went public in December. That bakes in success in new markets that it has recently entered, including groceries and pet food, and implies that it can convince restaurants to pay it more to feature their wares high up in its app.

Real business flywheels do exist. Software makers have managed to lock users in and thus generate gross margins typically above 70%. Lured by the promise of riches, venture capitalists are hoping against all hope to find news ones. They are already pouring money into the next generation of flywheel contenders: instant-delivery startups, which offer gratification in 30 minutes or less. Coupon-collecting consumers in cities such as New York now get at least a week’s worth of groceries for nothing from such services as Buyk, Fridge No More and Gopuff. Eventually, these firms’ champions promise, their economics will be far better than those of an Uber or a DoorDash. In the flywheel economy hope and hype spring eternal, at least as long as interest rates remain low and capital is essentially free.