Smaller companies looking to profit from artificial intelligence (AI) face an uphill struggle in China due to competition from the established big internet names and a more miserly funding environment, delegates at different conferences heard earlier this month in Hong Kong.
But there are coping strategies open to AI startups that could yet, in time, help them hit the big time, industry insiders said.
In a survey at the Asian Financial Forum (AFF) in Hong Kong, 31% of participants said bigger established players – the likes of Baidu, Alibaba, and Tencent (BAT) – were the number one threat to AI startups in China.
“Everyone is thinking about how to survive and thrive under such competition,” Bill Liu, founder of flexible screen maker Royole Corporation, said during one of the panel discussions.
The BATs enjoy huge advantages when it comes to developing AI because of the huge amounts of data they collect and control through their various online platforms. This is because data forms the essential building blocks of AI training.
In contrast, AI startups have to do it differently. One alternative method is to pay third parties for their data; another is to hire people to complete multiple-choice tests online.
Jason Tu, founder of fintech firm MioTech, said his company uses both these methods; speaking at a second forum hosted by S&P Global Intelligence, he said they were relatively expensive but potentially effective ways for smaller companies to build up their AI research capabilities.
Getting funding for such endeavours is getting harder, nonetheless. According to data from research firm ITjuzi, there were 602 fundraising events in the Chinese AI sector in 2018 – down 20% on the previous year.
Tu said he had sensed a hardening in investor attitudes in the second half of last year, with more of an emphasis on how and when AI startups would actually make money.
For investors to focus on the potential commercialisation of AI products is only right and proper – this is business after all, not charity. But given the huge longer-term potential of AI technologies, some investor patience is also merited, speakers said.
“It is understandable,” Liu said, “but [they] also need to understand that we are still three to five years ahead of AI adoption and startups need time to grow.”
So how to square that circle? By focusing on what might have the best chances of triumphing eventually.
“Many startups are bound by the thinking [about] how to make fast money,” Max Yuan, founder of Xiaoi Robot Technology, said during one of the AFF panel discussions. “Startups need to have irreplaceable technology innovation so they can make money in the long term.”
It follows that those AI-driven products that are hardest to replicate will most likely thrive and survive the longest, assuming they meet a clear need.
Finding that requisite X factor is the toughest nut to crack but Yu Cheng, a partner at Morningside Venture Capital, has some advice.
In an interview with FinanceAsia on the sidelines of a third conference, he said AI startups needed to figure out the various possible applications for their products before looking to plough more money in.
Having clear goals would help AI companies find their niche in the market, he said.
It might also help them to attract more funding.