Specialist
Former Sales Manager at SenseTime
Agenda
- US’s investment restrictions on SenseTime – background and impacts
- SenseTime Technology’s main competitors across computer vision segments, AI software and hardware product and service capabilities, advantages and disadvantages
- SenseTime Technology’s development strategies and investment focuses across vertical segments, business development strategies, revenue growth trends and drivers
- Smart Business, Smart City and Smart Life business AI platforms and application features, problems addressed, plus core technical competencies and pain points
- Typical project profit model, plus reasons for SenseTime Technology’s increasing gross profit 6. Terminal client and system integrator client cooperation model – sales strategies and process, product delivery and unit economics
Questions
1.
The US Department of the Treasury added SenseTime Technology (SenseTime) to a list of “Chinese military-industrial complex companies” on 10 December. According to the corresponding policies in the US, once a company is included in the blacklist, US investors are not allowed to invest in it anymore. Could you share your understanding of the news? How would SenseTime be influenced and what are the company’s countermeasures?
2.
According to Frost & Sullivan’s report referred to in SenseTime’s prospectus, in terms of the 2021 revenue scale, SenseTime has become the largest AI software supplier in Asia and the largest computer vision software supplier in China, holding an 11% share of the Chinese market. It has products that can be applied by businesses, city management, individuals and vehicles. What do you think are SenseTime’s competitors in the four sectors respectively? How competitive is SenseTime in these sectors?
3.
SenseTime’s revenue is mainly driven by business clients and city management projects. The revenue contributions by individual customers and the auto industry are limited. You have mentioned that you think the company has a promising outlook by supplying auto manufacturers. Could you further explain that?
4.
What are SenseTime’s advantages and disadvantages in its Smart Business and Smart City AI software and hardware products, technical capabilities as well as resources compared with Hikvision, Dahua, Tencent, Baidu, Huawei as well as the other three of the four AI unicorns, namely Megvii, CloudWalk Technology and Yitu Technology?
5.
Could you further introduce the competition in the security sector to work on provincial-level as well as prefecture-level or county-level projects? What is SenseTime’s development potential in these markets respectively? Could you also talk about SenseTime’s current penetration rates as well as development potential? Among Hikvision, Dahua and Uniview, which company may gain the biggest market share in the future?
6.
SenseTime’s revenue was RMB 1.85bn in 2018, RMB 3.026bn in 2019, RMB 3.446bn in 2020 and RMB 1.651bn in H1 2021. It seems that the company’s revenue growth slowed down in 2020 and was still not as rapid as before in 2021. What do you think were the major reasons? Which business sector do you think the company has delivered a sluggish performance in?
7.
Could you introduce the supply and demand trends in the Smart Business sector and SenseTime’s positioning and strategies in this sector? It seems that the revenue growth in this sector has slowed down. Why?
8.
On the enterprise side, what scenarios have high requirements on precision and software algorithm capabilities? What are the advantageous fields of SenseTime and other AI technology enterprises? How do you think of the application scenario’s future development?
9.
Let’s focus on the consumption application scenario. The revenue proportion of SenseTime’s Smart Auto and Smart Life segments accounts for 10% of the total. In H1 2021, SenseTime’s revenue in the Smart Auto and Smart Life segments reached RMB 70.6m and RMB 147.8m, respectively. In the consumption market highlighting mobile phone applications, ArcSoft is one of SenseTime’s competitors. There are also various competing solutions in the Smart Auto segment. What is the outlook for SenseTime in the two segments? What advantages does it have in the future smart auto market?
10.
Let’s focus on SenseTime’s specific products and overall economic models. According to SenseTime’s prospectus, it depends on SenseCore in developing and delivering standard software platforms with modular flexibility to clients. SenseCore offers a smooth, standardised and end-to-end automatic production flow of AI models. Based on SenseCore, SenseTime develops SenseFoundry-Enterprise for Smart Business, SenseFoundry for Smart City, SenseME, SenseMARS and SenseCare for Smart Life and SenseAuto for Smart Auto. Can you introduce the technical capabilities of SenseCore and the four platforms? What are their relationships? How do you compare SenseTime with its peers regarding AI-platform products?
11.
What measures can SenseTime take if it needs to deliver AI-empowered solutions to a business client but lacks related data? Will it be authorised to use the business client’s data?
12.
What do you think of the foundation model concept claimed by SenseTime? Do we just build this model with many small models like building blocks and apply this model to actual scenarios? Or can this foundation model directly enhance the production efficiency of specific small models?
13.
How replicable are SenseTime’s solutions to the Smart Business and Smart City scenarios? What is the proportion of customised solutions in these scenarios? What are the pain points involved in implementing customised solutions?
14.
To make engineered products and specific solutions reusable, what is the other content SenseTime needs to add?
15.
How would you explain SenseTime’s constantly increasing gross profit? In theory, the more customised solutions it offers, the higher implementation costs SenseTime will bear. However, SenseTime’s gross profit margin increased from 56% in 2018 gradually to 70% in 2020 and then to 73% in 2021. SenseTime enjoys a gross profit margin that is much higher than that of most AI companies. Why?
16.
We noticed that SenseTime and many other AI algorithm companies are constantly increasing their R&D investments. Do you think it’s necessary for them to do so? If so, what’s the reason? If it isn’t, can these companies start to turn a profit after stopping their R&D investments or laying off some employees?
17.
SenseTime includes terminal clients and system integrators in its overall client group. This company doesn’t view system integrators as its distributors, and instead, it still treats these integrators as they are part of its clients. Can you compare the models under which SenseTime cooperates with the two types of clients? Especially, how does SenseTime participate in government projects?
18.
Why did SenseTime set an account receivable turnover period of as long as 490 days for government clients?
19.
Can you introduce the economics of a typical project undertaken by SenseTime, including the solutions, charging models, software and hardware deployed as well as the total contract amount? After the delivery of software and hardware in the first year, does SenseTime have any subsequent revenue source, say, the offering of services?
