AVCJ Awards 2017: Deal of the Year - Early Stage Technology: SenseTime

Xiao'ou Tang, a professor in the department of information engineering at the Chinese University of Hong Kong (CUHK), is an expert in facial recognition technology. Over the course of more than 20 years, he guided hundreds of PhD students conducting research in computer vision and pattern recognition, but little progress was made in terms of commercialization. The technology just wasn't mature enough.

The turning point came in 2011. Since then, the development of deep learning – which involves using neural networks to feed data into a computer, which is then able to make decisions – has accelerated. And it has taken computer vision with it, is essentially enabling a range of computer vision applications, from face and image recognition to machine vision for self-driving cars.

Between 2011 and 2013, Tang and his students published more than 14 research papers on deep learning in computer vision – out of 29 papers in this field globally. As a result, Tang recognized an opportunity to commercialize the technology to which he has devoted his career. He established SenseTime three years ago in Hong Kong Science Park, alongside two CUHK PhD students – Li Xu and Bing Xu – and Fan Yang, formerly a software development engineer at Microsoft.

"In China, there haven't been many start-ups formed by teams of professors and university students who want to commercialize their scientific research, although this is commonly seen in the US," says Bing Xu. "Most Chinese companies are focused on the innovation of business models, but I expect there will be more technology innovations in China, led by university spin-outs."

Infrastructure build-out

IDG Capital visited the CUHK campus for due diligence in late 2014, and soon invested $10 million in a Series A round. Within a year, SenseTime had expanded its four-strong founding team to include 30 scientists and industry experts, most of whom are former students of Tang.

The commercialization process was expected to take two years or more and require a huge upfront capital commitment to hire scientists and purchase equipment so the company could build internal computation capability. However, few Chinese VC firms foresaw the contribution computer vision and deep learning technology would make to the development of artificial intelligence (AI) software, an area that attracted a lot of investor attention in the past year.

"In 2014-2015, we talked with many VC firms, but they only focused on our revenue. We were burning cash to recruit talent and build a proprietary deep learning platform at that time," recalls Xu. "We went through some tough times. On several occasions, we could hardly pay our employees." Tang sold his apartment and dipped into his savings to keep the business afloat as well as obtaining about RMB50 million in venture debt funding from Silicon Valley Bank.

SenseTime wants to build a proprietary deep learning platform – known as Parrots – that drives core AI technology, much like Google's TensorFlow. The company has already developed more than 5,000 different types of graphics processing units (GPU), a hardware that can accelerate the performance of deep learning. It hopes to help reduce R&D costs for AI technology developers in a variety of industries including medical imaging, high-performance computer chips and autonomous driving.

"We think the AI industry will achieve an explosive growth in the future. To achieve that you need to build a solid computing infrastructure to process a large amount of data. It's similar to the internet industry – first thing you need to have a comprehensive broadband network coverage," says Xu. "Our infrastructure coverage is probably more than four years ahead of that of other new players in the AI industry."

The victory of Google's DeepMind System, dubbed AlphaGo, over a human professional Go player without handicaps in October 2015 is regarding as a breakthrough for the AI industry. Since then, financial investors have shown greater interest in deep learning and AI technology. For SenseTime, AlphaGo's win represented a maturation of deep learning technology, and the company responded by looking for new funding to scale up its business.

A start-up in demand

SenseTime received approximately $10 million in a round led by StarVC, a VC firm established by several Chinese movie and entertainment personalities, in April of last year.

Although the company is not yet profitable, it supplies its AI chips and software solutions to the likes of China Mobile, Huawei, Xiaomi and JD.com. This was enough to convince CDH Investments of its commercialization potential. The PE firm led the first tranche of a Series B round worth $120 million in December 2016, joined by Dalian Wanda Group, IDG and StarVC. A second tranche of $290 million, led by Sailing Capital, closed six months later, valuing SenseTime at more than $1 billion.

"Many investors were surprised that CDH, a private equity firm that has historically invested in traditional industries, would actually back a technology company that has not yet generated any profit. Given CDH's strong reputation, many other PE firms were confident enough to invest," says Xu. 

Indeed, there were as many as 20 additional participants in the Series B round, including China Merchants Securities International, Morningside Venture Capital, China International Capital Corporation, Co-Store Capital, Advantech Capital, and China Renaissance.

SenseTime is now worth in excess of $2 billion, having leveraged its investors' commercial and government relationships to accelerate growth. The company has more than 400 customers in China and overseas and last month announced a tie-up with Qualcomm Technologies to collaborate on AI and machine learning for mobile and internet-of-things (IoT) products.

"If we launched our business in the US, we would likely get acquired by large tech companies, like Google or Facebook, at an early stage of development," says Xu. "In China, however, large tech firms are not so keen on acquiring core technology companies. That's probably one of the reasons why we've become a major AI player in China." 

Pictured: Stuart Schonberger of CDH Investments

• Tweet  
• Facebook  
• LinkedIn  
• Google plus  
• Save this article  
• Send to  

• Topics
• Greater China
• Early-stage
• Technology
• China
• Venture
• CDH Investments Management
• IDG Capital Partners
• AVCJ Awards