“Are there areas where you see there is too much hype?” Elzio Barreto, a Reuters correspondent asks three veteran China-focused venture capitalists on stage at RISE Conference in Hong Kong. He cites a recent incident where an umbrella sharing company in Nanchang, the capital city of Guangxi Province, lost all its 30,000 umbrellas within a few weeks from its launch. As the investors chuckle, they come to the consensus that China’s sharing economy is overhyped.

“The sharing economy, the entire thing where you can scan a QR code and pay one RMB for something, is being overly hyped up in China,” says Harry Man, Partner at Matrix Partners China. Man oversees investments for the China branch of Matrix, whose notable early-stage investments include Apple and SanDisk.

In the last few years, the “sharing economy” has become a warm and fuzzy word for Chinese entrepreneurs who dream of facilitating the community sharing of everything, from umbrellas, orange juice machines, stand-alone KTV booths, massage chairs to the most recent buzz around power banks. At the beginning of this April, five power bank rental startups received a combined RMB 300 million ($43 million) financing within ten days with 20 investment firms involved. The Chinese government has also jumped on the sharing trend, expecting the sector to grow about 40% this year to 4.83 trillion yuan ($705 billion) and account for around one-tenth of GDP by 2020. Most of these sharing promises, however, will never take off to the extent that car and bike sharing did.

The most successful startups under the sharing economy umbrella are those that have network effects and barriers of entry, states Helen Wong, Partner of Qiming Venture, an early investor in China’s popular bike rental company Mobike whose 3 million MAU (in Chinese) have probably inspired the boom of the sector. “The last mile transportation in China is still an issue, so bike sharing is a good way to solve that problem,” she adds. China also has a big number of densely populated cities, where 128 cities have more than a million population, so the demand is big enough.

“We are a believer in bike sharing, but we are not sure about umbrellas or power chargers,” says Wong.

“For VCs, we just need to look into the real numbers,” Man concurs. “Whether there is a big enough market for them to sustain at least for us a multi-billion-dollar company creation.” Instead of platforms that would inspire the collective good and create less waste, a lot of startups in this space have become, Man argues, purely rental businesses “. . . which are workable, but not sexy.”

Another rising trend that has made entrepreneurs and investors excited is artificial intelligence.

“AI is the next mobile,” claims Hans Tung, Managing Partner at GGV, affirming that AI has a real demand and the potential to take off. According to a report put out by NetEase and Wuzhen Think Tank (in Chinese), in 2016, China came second in terms of the amount of investment in AI companies on a total of $2.57 billion, after the US which invested seven times as much. In Q4 of 2016 alone China saw 173 investment deals in AI, a record high season in the past five years.

“AI is going to be happening. It will happen like what happened to PC and the internet. It will trickle down into every vertical industry. But how fast would that be and how much opportunities are there going to be created is still waiting to be seen,” says Man. According to the NetEase-Wuzhen report, 81.4% of the companies are early stage (in Chinese), which means the AI industry is in its infancy.

“How AI is different from mobile is that there is the whole B2B area like customer relationship management,” Man observes. “There are fewer opportunities on the B2C side and even when there are, the big companies like Alibaba, Tencent and Baidu are going to have an advantage.” Last week, Baidu was speeding ahead with its first AI conference in Beijing where it announced an open platform for autonomous cars.

How do small startups compete with the big companies in AI? Tung provided the solution with the example of a GGV-backed startup. Liulishuo, powered by voice recognition and AI, helps people speak accent-free English by recording what the learners speak, build machine learning on top of that voice data, and give tips on improving pronunciations. Today the app has a total of 45 million registered users. The key for a small startup to succeed concludes Tung, is to find specific, targeted areas where startups can bring innovation with a strong, creative team.

Tung observes that startups are spending too much time on building AI infrastructure. “Whoever is creating the internet doesn’t make any money.” The ones who are going to make billions of dollars, he argues, are the ones who are “creating an application on top of the internet and is disrupting the market.”