Is USD 7 Billion an Overestimation for SenseTime?
Entering the commercialization period of the AI industry, SenseTime has to prove its profitability instead of leaning on the AI concept and academic research.
This article is translated from iyiou (https://www.iyiou.com/p/111617.html) by Zhou Wenmeng.
Li Xu, the CEO of SenseTime, in a speech in Singapore on September 5th, 2019, claimed that the estimated value of SenseTime had exceeded USD 7 billion.
According to CB Insights, SenseTime had reached USD 4.5 billion estimated value in 2018, after receiving funding from institutional investors including Alibaba and IDG Capital.
So does the USD 2.5 billion increment make sense?
From facial recognition startup to CV unicorn
Founded in November 2014 with DeepID, a facial recognition algorithm with a 98.52% accuracy rate – exceeding Facebook’s DeepFace algorithm at the same time period – SenseTime has been chased after by VC Capital firms ever since.
With ample funding from top institutional investors, SenseTime has extended from facial recognition to more professional Computer Vision (CV) technologies, including large-scale video data understanding/mining and image processing /augmentation. Based on such technologies, SenseTime has become the AI algorithm provider of industrial giants in autonomous vehicle, healthcare, smart city, etc.
Is USD 7 billion an exaggeration?
SenseTime is now facing great challenges as an AI algorithm provider, although with the environmental shift that AI is now a relatively mature and commercialized industry.
There are mainly three different types of companies in the Chinese AI industry – algorithm provider, AI application developer and companies designing industrial upgrading solutions. The first type of company, like SenseTime, has weaker pricing power compared with traditional industrial giants, acting in the second or third roles. The open source characteristic of AI algorithms research puts most algorithm provider companies at the same research level. And the quantity and quality of data is more valuable than cutting-edge AI models.
Industrial giants, with massive data, are therefore more decisive in comparison. So SenseTime needs to build tighter connections with the industries.
It is hard for a single company to provide AI-related services for companies in several industries as the commercialization of AI technology requires deep processing of industry-specific data. SenseTime is, however, landing in a wide range of industries. Yet the scenario whereby SenseTime receives high revenue by widely providing AI technology and then pushing its estimated value to a higher level will no longer exist in the near future, with the recent downward-trending economy will mean companies are more budgeted in application of new technologies.
The rising USD 2.5 billion in estimated value of SenseTime is, as such, very doubtful.
The good news is that SenseTime is having more collaboration with the semiconductor industry. It has developed an AI chip that is likely to be the complement of NVIDIA’s products. And SenseTime may receive more support from its strategic investor, Qualcomm Incorporated.
AI chip making, a sector that is highly promoted by the Chinese government, can provide a chance for SenseTime to release its capability in AI technology.
Postponed IPO plan
News also came out that SenseTime did not have an IPO plan in the near future and was more interested in raising PE funding.
SenseTime is now doing roadshows around the world. Its plan is to go IPO overseas after receiving ample funding from the domestic PE market first. Entering the commercialization period of the AI industry, SenseTime has to prove its profitability instead of leaning on the AI concept and academic research. In conclusion, SenseTime, as part of the Chinese AI ‘national team,’ should realize its value by sharpening its AI technology for realizable industrial usage, which is much more important than its IPO plan, estimated value and social footprint.