Pony Ma's Tencent has joined a group of venture capital firms to invest in a Canadian artificial intelligence lab, as the Chinese social media and technology giant takes on domestic rivals in AI development.
Montreal-based Element AI said it had raised $102 million in the series A funding round. Besides Tencent, investors include South Korea's Hanwha Investment, the venture capital arm of the Hanwha conglomerate.
Lead investor was Data Collective (DCVC), while Fidelity Investments Canada, Intel Capital, Microsoft Ventures, Nvidia and Real Ventures, also joined the fundraising round, according to a statement released on Tuesday.
Tencent's decision to invest is the latest move in a fierce competition among China's tech giants to be first to harness AI as a tool to monetise the vast quantities of data they have on record, analysts said.
Element AI, founded by a group including pioneering AI researcher Yoshua Bengio, will use the money to encourage businesses to take up AI and will create 250 jobs by January. It plans to open offices in Japan and Singapore, it said in a statement.
Element says its goal is to help companies solve problems by using AI technologies.
"Artificial Intelligence is a 'must have' capability for global companies," Element AI CEO Jean-François Gagné said in the statement. "Without it, they are competitively impaired if not at grave risk of being obsoleted in place.”
Chinese technology giants are gearing up their expansion in the AI sector because they have collected too much data they cannot fully commercialise, said Connie Gu, a Beijing-based analyst at Bocom International.
“Every company is trying to figure out how to monetise their data in a larger scale,” Gu said. “There’s tons of data out there, but the problem is how to use them. [The Chinese tech giants] are still lacking in the ability to process data, selecting the useful ones and putting them into their businesses. Right now only a fraction of data is being used.”
Among the technology titans in China, Baidu is recognised as the pioneer in the AI field. It was the first among domestic rivals to open an AI lab, in Silicon Valley in 2014. According to a Bloomberg report, the company spent over $2.9 billion on research and development in the past two and half years, most of which has gone on AI.
However, its AI division is facing management issues after chief scientist Andrew Ng, a Stanford University academic who helped the company make its foray into deep learning technology, announced on his Twitter feed that he was resigning. And that potentially leaves room for Tencent, which sees opportunities in the combination of AI technologies and its social mobile apps – such as its massively popular messaging-to-online payments app WeChat.
Gu said although Tencent's expansion to the AI sector was relatively late, it was catching up fast and exploring a broader scope through both investment and its own development.
Apart from Element AI, Tencent also joined a seed round for California-based deep learning firm Skymind last year. A program developed by the company to play the traditional Asian board game Go, Jueyi (FineArt in English) this year won the 10th Computer Go UEC Cup in Japan – although it has yet to face Google's Alpha Go technology.
"Tencent is actively pursuing investment opportunities in AI and innovation globally," Hongwei Chen, Tencent's executive director of investment and M&A, said in the statement. "We are excited to partner with Element AI and Real Ventures, and look forward to seeing breakthroughs in AI technology and applications that help improve people's lives."
In April, Tencent established its first US-based AI research lab in Seattle, dedicated to providing AI technical support for mobile products and services.
The lab is led by former Microsoft scientist Yu Dong. Yu is expected to lead Tencent’s research on speech recognition and natural language understanding, a useful technology for Tencent’s messengers and social network products.
“Before Tencent was focusing on (the AI technologies) that can support its own businesses, such as imagine analysis and speech recognition,” she said. “It had a late start, but a broader scope.”
