Bike-Sharing Collected $1.8 Billion in a Year. How Did Internet Companies Learn to Blow Bubbles This Fast?

The playbook for spawning a big internet company keeps getting more practiced, but nobody has found a way out yet

No unprofitable new business has ever pulled in this much capital this fast the way bike-sharing has.

On June 16, Mobike announced a $600 million round. That brings its total over the past 12 months to $1.1 billion.

Its biggest rival, ofo, said this week at the Summer Davos forum in Dalian that a new round is "close to completion." According to an earlier Reuters report, it will be roughly the same size as Mobike's. Over the past 12 months, ofo has raised $700 million.

Both companies are now valued at more than $2 billion. Each likes to celebrate closing a round by subsidizing users with free rides. We built a calendar of the free-ride days in Beijing:

At a Davos roundtable in Dalian yesterday, a CNBC reporter asked Mobike founder Hu Weiwei when the company would turn a profit. She answered with a scene from the film The Social Network.

In it, the newly minted entrepreneur Mark Zuckerberg meets the seasoned Sean Parker. Zuckerberg is hesitating over advertising; he worries that running ads will make Facebook uncool. Parker eggs him on: a million dollars isn't cool — wait until the company is big enough to make a billion.

David Fincher shot the scene beautifully, and it fits today's startup rhetoric perfectly: an internet company shouldn't think about making money at the start, it should get big first. By now this passes for gospel.

Except that the script of The Social Network is nonsense from beginning to end, and this scene is no exception.

In reality, Zuckerberg did take plenty of investment, but he was selling ads from his first year in business. Less than a year after launch, Facebook's ad revenue had passed the "uncool" $1 million mark, and from 2005 on it grew faster than the user base did.

By the time it listed on Nasdaq in 2012, Facebook had been profitable for three straight years.

By the time Facebook turned profitable, it had raised only $500 million — less than Mobike takes in a single round.

Mobike and ofo aren't just unprofitable. As free-ride days pile up, they don't even bring in much revenue. Users can win cash prizes just for riding.

And they are not alone. Didi and Uber have each raised more than $10 billion and are still burning cash. Food delivery, after several rounds of consolidation, is said to be burning cash in something called the "second half."

What, in under 20 years, let a Hollywood screenwriter's fantasy replace Zuckerberg's real story as the reality of internet entrepreneurship?

Burning cash to blow bubbles began the moment internet companies were born

An internet entrepreneur five years out of college started his second company: cat food, dog food and pet supplies delivered to your door.

Within six months, because it sat on top of a hot trend, his company raised $50 million in venture capital, and the biggest e-commerce company in the country took a stake.

Canned cat food is not light, and shipping it cost a lot. Neighborhood pet stores were already everywhere. But money came easily, so the company sold below cost and threw in free shipping.

To build recognition, it spent roughly 100 million yuan on expensive ad placements. The effort worked: 18 months after it was founded, the company went public, and its market value at one point reached $6 billion.

This is not the daydream of a Chinese O2O founder in 2015. It is the true story of pets.com, the pet-supplies site the American entrepreneur Greg McLemore founded in 1998.

Life after the IPO did not go well. Its financial disclosures show that from February to September 1999, pets.com spent $11.8 million on advertising, including $2 million to put its mascot in a 15-second Super Bowl spot that cost a fortune. Over the same months, its revenue was $619,000.

On November 6, 2000, hours after George W. Bush was elected president, pets.com announced it was shutting down. From founding to IPO to delisting, the whole run took 26 months.

pets.com was one of the stars of America's 2000 internet bubble. Most of the companies that died then followed the same arc: an idea with the internet in it → a few rounds of venture funding → burning cash to subsidize growth → an IPO → collapse.

E-commerce companies had it worst. Making up 10% of the Nasdaq index by weight, they were the epicenter of that year's crash. The E-Commerce Times Stock Index, which tracked e-commerce share prices, fell from a 1999 high of 340 to 61 in 2000, a drop of 82%.

A contemporary of pets.com, a company called Kozmo, promised to deliver food, DVDs and convenience-store goods within an hour, and later added high-priced branded goods. A business like that could not survive even today, let alone in 2000, when you had to sit at a desktop computer to get online.

There were stranger companies still. Kinko's, the self-service photo-printing chain, launched a $12-an-hour internet service that let customers sit in its stores and go online.

In March 2000, more than 6,000 companies were listed on Nasdaq, the great majority of them tech companies. They had three things in common: founded recently, listed quickly, losing money. The average price-to-earnings ratio on the Nasdaq index reached 175. If those companies kept earning what they earned that year, an investor putting in $1 would need 175 years to get it back.

When the bubble burst and share prices collapsed, most of them simply disappeared.

The biggest reason to blow bubbles: everyone involved could get money

During the 2000 internet bubble, the people inside it knew perfectly well that it was an enormous bubble.

Yahoo's founder Jerry Yang, for one, was uneasy after the company's 1996 IPO. He later recalled: "I remember driving home, and driving back to the office, and thinking, 'Oh my God, what have I done.'"

Yahoo was then an index of web pages — what would later be called a portal — and it made its money from advertising. In 1995 it took $2 million in venture funding from Sequoia Capital to keep expanding.

Judged on revenue and profit, Yahoo was in poor shape when it listed. Before the IPO it had $1.4 million in ad revenue and a $643,000 loss. On numbers like that, Yahoo's market value topped $100 billion at the peak of the bubble. It proved a few years after the bubble burst that it was a decent business, but in the late 1990s it was plainly overvalued.

Its investors, Sequoia and Masayoshi Son, presumably did not share Yang's unease. On Yahoo's first day of trading, the value of Sequoia's stake rose 8,500%.

Over the previous 20 years, the average annual return on tech venture capital had been 15.4%.

Venture investors could cash out once they had pushed a company onto the public market. That is the main reason every sort of shaky company could raise money then.

The investment banks that underwrite shares during an IPO were supposed to be wary of shaky companies, since they had to be sure the stock would sell once it listed.

But in that market, any company connected to the internet drew a crowd of buyers the moment it listed. As the New York Times put it at the time, every American household had a stock-tip pamphlet lying around and everyone had "B2B e-commerce" on their lips.

So the underwriters, the gate a company had to pass to list, quickly relaxed their standards. Credit Suisse dropped its profitability requirement and replaced it with a rule that a company needed at least $10 million in revenue in the 12 months before listing — and then stopped insisting on that, too.

Taking one company public could earn a bank as much as $500 million.

So although the banks and the venture firms could see everything wrong with those internet companies, self-interest led them to paint a beautiful future, sell the shares to as many people as possible, and take the commissions.

As for the people buying frantically after the listings, they probably did not believe in the companies either. They were afraid of missing out while prices rose so fast.

Over the past few centuries, nearly every bubble has formed because the people in it believed they could make their money before it popped. Even Isaac Newton, who discovered the law of universal gravitation, got caught in the South Sea Company bubble.

As Benjamin Graham, Warren Buffett's teacher, wrote in The Intelligent Investor: "As early as the spring of 1720, Sir Isaac Newton owned shares in the South Sea Company, the hottest stock in England. Sensing that the market was getting out of hand, the great physicist muttered that he could calculate the motions of the heavenly bodies, but not the madness of the people. So Newton sold his South Sea shares, pocketing a 100% profit totaling £7,000. But just months later, swept up in the wild enthusiasm of the market, Newton jumped back in at a much higher price — and lost £20,000 (equivalent to roughly $3 million in 2002-2003 dollars)."

China's internet bubble probably began with group buying in 2010

China had internet companies of its own during America's bubble — 8848, NetEase, Sina and others.

But unprofitable companies could not list on the A-share market, and the securities regulator also limited overseas listings. By the time Sina used a VIE structure to list in the United States, it was early 2000 and the bull market was nearly over. They missed Wall Street's frenzy.

NetEase, which listed that August, ran straight into the bursting bubble; its shares fell below the offer price on the first day.

China's internet bubble really starts with group buying.

Groupon invented group buying in 2008: the site used discounts to push users toward restaurants, karaoke parlors and cinemas, and took its money from those venues.

Group-buying sites had large transaction volumes and low costs. But the United States was still recovering from the internet crash and had run into the financial crisis on top of it, so capital did not flow to startups like these.

Then Groupon gathered 85 million users in 2010 and reached a valuation of $1.35 billion, and its imitators came up all over China.

The most successful pair of rivals, Lashou and Meituan, both founded in 2010, raised $60 million and $12 million respectively in their first year, then $110 million and $50 million in their second. Numbers that look ordinary today were rare and very large then.

With that money in hand, China's big group-buying companies all prepared to burn hundreds of millions of yuan on advertising.

• Lashou put most of its new $50 million round into advertising;
• Nuomi said it would spend 200 million yuan on advertising;
• Meituan announced a 130-million-yuan advertising tender;
• Manzuo said it would throw 100 million yuan at advertising;
• Dianping said 200 to 300 million yuan of its $100 million round would go to advertising.

According to the Annual Report on China's Group-Buying Industry from the China E-Commerce Research Center, the number of group-buying sites in the country peaked in mid-2011 at close to 5,200.

Their business models were almost identical: start with restaurants, send street-level sales teams to sign merchants up and negotiate a cheap set menu; meanwhile advertise everywhere to draw users to order on the site, then split the money with the merchant as agreed.

Lashou once prepared to list in the United States with a Groupon-style prospectus of heavy losses and fast growth. It withdrew the application in 2012, and its valuation shrank from $1.1 billion to $600 million. The failed listing, and funding that could not keep up, led directly to Lashou's sale to Sanpower Group.

By 2014 only 176 group-buying sites were left in China, a survival rate under 4% from the peak. What remained was essentially the merged Meituan-Dianping and Nuomi, which Baidu had bought.

A wave of O2O arrives, and fake orders join subsidies as the new trick

Group buying pulled people from the web into stores.

O2O, which took off in 2014, pulled services out of stores and into people's homes. These companies used subsidies to get users to summon any service they liked to their door with one tap.

Helijia let users pick a manicurist and a time slot in an app, place an order, and wait for the manicurist to arrive — exactly like ordering food delivery. In July 2014, IDG put 30 million yuan into Helijia's Series A.

Three months later came a Series B of tens of millions of dollars from Broadband Capital and IDG Capital, bringing Helijia's valuation to 1 billion yuan. From zero to a billion in six months.

"The VC money I've got is enough to burn for years. What's the rush? First expand the scale, take the market. Manicures are an industry I intend to monopolize." That is what founder Meng Xing said at the time.

Gongfuxiong, founded in 2014, at first let users call a masseur to their home for as little as 66 yuan — cheaper than a massage parlor, mainly because of cash subsidies. Within a year of its founding, Gongfuxiong had raised $13 million.

According to the British research firm Preqin, by December 2, 2015, the total value of venture capital deals in China had passed $34 billion, double the 2014 figure.

In 2015 the People's Daily called Gongfuxiong an "industry leader."

That year the People's Daily said other things too — that 4,000 points, for instance, was the starting point of China's bull market.

That July, shaken investor confidence wiped $5 trillion off China's stock market. Gongfuxiong, likewise, never raised another round.

The star companies of that moment have gone almost completely quiet.

Subsidizing users was expensive, so some O2O companies simply faked orders to manufacture their numbers. Aishang Xianhua was one. By the figures it disclosed itself when it listed on the National Equities Exchange and Quotations, the "New Third Board," 42% of its orders from January to July 2015 were its own fakes.

Data compiled by Bloomberg shows the value of Chinese M&A deals jumped 80% in 2015 to $515.6 billion, while domestic internet-industry deals grew fourfold year on year to $67.4 billion.

Helijia's rival Doudou Meijia folded in February 2016 and sold itself cheap to 58 Daojia. 58 Daojia CEO Chen Xiaohua told Securities Times: "Since last year a huge number of O2O companies have died. More than 10 manicure O2O companies alone have come to us hoping to sell themselves."

O2O startups appeared in a wave over a short span — beauty, massage, car washes, medicine delivery — as though nothing were beyond them. In 2015 a QDaily reporter ran an experiment: 18 summer days at home without going out, trying the services of 72 companies.

A year later only one in seven of them had raised follow-on funding. The rest had shut down, pivoted, pivoted and then shut down, or quietly vanished from users' view.

Of everything delivered to your door, only food delivery still has money to burn

Meituan and Dianping, survivors of the war of a thousand group-buying sites, each opened a food-delivery channel in early 2014. Meituan attracted $1 billion of investment that year; Dianping took more than $100 million from Tencent, then used part of it to invest in Ele.me.

Before that, Ele.me looked about as small as China's food-delivery market itself. The company had launched mobile web ordering back in 2010, but as late as early 2014 it still had only about 200 employees, covered 20 major cities, and had raised a few million dollars.

Only after the $80 million it took in 2014 from Dianping and its existing investors did Ele.me enter a period of frantic growth: a huge street-sales operation expanded it to more than 250 cities and 200,000 merchants within a year; it plugged in more third-party logistics teams, pushing daily orders past one million and active users past 20 million.

At the end of 2015 Ele.me signed an investment framework agreement with Alibaba and received $1.25 billion from it. That took Ele.me's valuation straight to $4.5 billion.

The subsidy habit that began with group buying carried into food delivery. Ele.me and Meituan do not run restaurants; they raise money again and again to fill the subsidy hole, and that money goes to merchants, users and delivery riders. Open a food app and "10 yuan off orders over 25 yuan paid online" or "8 yuan off your first order" is sitting on the home page.

The research firm Hudongpai's August 2015 Food-Delivery O2O Survey Report found that when Ele.me canceled its "15 yuan off orders over 50 yuan" promotion that February, users moved quickly to Meituan Waimai, which was discounting more heavily. When Ele.me brought the promotion back, more users came back to order from it each day.

Burning cash on services that are all alike makes customers hard to hold on to. So even now, with only Meituan Waimai, Ele.me and Baidu Waimai left, and with the market supposedly in the "second half" that follows all that cash-burning, Meituan still drops a 4-yuan cash red packet on every order, and Ele.me varies its subsidy by time of day.

The biggest of all is ride-hailing

Didi and Kuaidi were both founded in 2012. In their first year Didi raised $20 million and Kuaidi $10 million. They began as taxi-dispatch platforms, able to reach at most ten thousand or so drivers in any one city.

From 2014 the two companies — nearly identical in business model, staffing and core markets — raised money faster and in bigger amounts, each trying to beat the other with subsidies.

By August 2014 Didi and Kuaidi had spent more than 2.4 billion yuan on subsidies, some to drivers, some to riders. Wang Gang, an early Didi investor, later recalled that Tencent had budgeted only one week of subsidies for the fight, but Alibaba, backing Kuaidi and fighting for control of the payments gateway, kept paying more, and neither side could stop.

Not until February 2015 did Didi Dache and Kuaidi Dache issue a joint statement announcing their merger. That ended the first round of the ride-hailing war, the one fought over taxi drivers. In 2016 Didi founder and CEO Cheng Wei looked back on the subsidy war: "We started with a 4-million-yuan budget and never imagined we'd spend more than 100 million that month. My hand was shaking when I signed off on it." Between them, the two sides had raised nearly $1.5 billion by then.

A second subsidy war began immediately afterward, this time between Didi and Uber. What they were competing over was the ability to mobilize millions of private-car drivers.

The new story drew new money. Piling up capital quickly gave Didi the means to fight Uber subsidy for subsidy:

• May 2015: Didi Kuaidi raises $142 million in its Series E
• July 2015: Didi Kuaidi raises $3 billion in its Series F
• February 2016: Didi Kuaidi, now renamed Didi Chuxing, raises another $1 billion in its Series F
• May 2016: Didi Chuxing raises $1 billion in strategic funding
• June 2016: Didi Chuxing raises another $1 billion in strategic funding

In 2015 alone, the two companies at one point handed out ride subsidies at a rate of more than $1 billion a year. Much of it went to driver incentives.

At the peak of the war, Yidao came in carrying a large pile of capital. Its subsidy was blunter: top up your account and get 100% of it back. By Yidao's own later account, that promotion ran 227 days, during which users topped up 6 billion yuan and Yidao paid back 6 billion yuan — more than LeTV's net profit for a year.

But next to Didi and Kuaidi, Yidao was still too small, and even an enormous subsidy could not lift its market share quickly.

In August 2016 Didi China issued a statement confirming that it had bought Uber China; after the deal Didi was valued at $35 billion. Add the $5.5 billion it took this past April from SoftBank, Silver Lake and others, and Didi's valuation passed $40 billion. A complete lock on China's ride-hailing market.

The trouble is that the $40 billion rests on capital's expectation that Didi will come to own the whole transportation business. Under the new ride-hailing rules taking effect city by city, both Didi's effort to overturn the taxi industry and the market's original expectations for it have evaporated.

Still, the subsidies, the brawling and the eventual merger of Didi and Kuaidi became the template for the startups that came next.

Spawning a big internet company keeps getting more practiced, but nobody has found a way out yet

Compared with the internet bubble 20 years ago, today's market is a different world. Back then the whole United States had 100 million internet users, most of them on desktop computers. Today China alone has nearly 700 million, going online through smartphones they never put down.

Just as Amazon and Alibaba both survived from 20 years ago, food delivery, bikes and ride-hailing will all stay in our lives.

The question is what the companies that survive are actually worth.

The method for quickly spawning companies is no different across the two eras: use subsidies to get as many users as possible onto your service as fast as possible, and count on finding a way to make money once the user base is big enough.

In the seven years since group buying, the method hasn't changed much; only the form and the scale of the subsidies differ. Group buying mainly subsidized users. Food delivery had to subsidize users and riders. Ride-hailing had to subsidize drivers and passengers.

The selection here is brutal. It is no accident that Meituan, having lived through several bubbles, went from group buying to food delivery and has now taken up chauffeured cars and home-sharing. Only the companies that get and spend resources more efficiently make it to the end.

In the end the few biggest companies all have enough capital, and with nothing to tell them apart, the burning ends in a merger.

But once they merge and own a market, does the new business make money — enough to justify everything poured into it first?

So far ride-hailing, group buying, food delivery and bike-sharing have all yet to prove they are still good businesses without subsidies.

Over seven years, not one of the industries spawned this way has found a way out, but the early growth has gotten more and more practiced.

Group buying and O2O each had hundreds of companies raising money. By the time of ride-hailing and food delivery, fewer than 10 could keep getting capital behind them.

By bike-sharing, after Mobike came to Beijing in the summer of 2016 and caught on, the two biggest companies each raised hundreds of millions of dollars in under two months, leaving every other rival far behind.

Capital locks onto the top two faster now, and less of it goes to anyone else.

As for why it is always two companies, that is probably no accident: the pool of giants able to take them over has itself shrunk, from Baidu, Tencent and Alibaba to just Tencent and Alibaba.

In all these years, few of these new companies have become sustainable businesses, but the cash-burning wars have guaranteed that they keep raising money, which also gives early investors a chance to sell their shares.

And so early-stage money arrives earlier and gets bigger.

But what happens in the end when one company too big to fail after another is left standing there? Not many people seem to care.

In 2015, Sam Altman, president of Silicon Valley's biggest incubator, Y Combinator, came to China with his friend Peter Thiel. That year Thiel's startup bible Zero to One was China's bestseller of the year, selling more than a million copies.

After he got home, Altman posted two things on Twitter:

"If discussions about tech always end up at 'are we in a bubble?', then a bubble probably isn't happening. What's worth worrying about is when everyone is wildly optimistic."

"Case in point: I was in China last week. It looked exactly like a tech bubble. Nobody was worried about it at all."

So it goes.

Cover image/BusinessInsider