The Internet of Things industry is harder than the IT industry. Because of the factors to be considered, more things to do, and less tolerance.
“The Internet of Things industry is harder than the IT industry. I have a say in this. In the whole process, you have to consider the factors and do more things. Moreover, the fault tolerance is smaller.”
This is the speech of Hikvision Senior Vice President Xu Ximing at the AMB meeting in an interview with 36 。. Compared with previous years, this year’s AMB will have fewer companies displaying chip PPT, and more companies showcased landing plans. The booths of Huawei, Dahua, Yushi, Yitu and Hikvision are very close, and the ability to display occasionally overlaps. Compared with several other competitors, Hikvision has a lot of low-key, no announcements, no billboards.
36 氪 As one of the two media involved in the interview with Xu Ximing, he discussed Hikvision’s new products, AI Cloud open platform and solutions in various industries. This year, Hikvision’s booth took a lot of space to showcase its software and hardware applications in different industries, involving public safety, transportation, manufacturing, retail, education, logistics and other industries. AI seems to be closer to life.
For example, in the field of food production, there is “Ming kitchen bright stove”; in the field of smart retail, there is “commodity identification”; in the field of natural resources, there is “sky recognition”.
In addition to the solution, Hikvision has also brought in new hardware – the radar video all-in-one. This is a series of RayVision products that incorporate radar and video technology. Among them, the Lei Shi road safety early warning integrated machine can realize the early warning of the collision of vehicles (or pedestrians) in two directions, such as level crossings, sharp turns, and parking lots.
For example, on a rural road with narrow roads and irregularly arranged houses, the driver has a blind spot due to the naked eye, and it is easy to have an accident when turning. The radar function of the Lei Shi Road Safety Early Warning Integrated Machine can detect danger in advance and link the LED screen for information broadcast. In addition, there are Lei Shi vehicle inspection device, Lei Shi event detection integrated machine, Lei Shi sequencing integrated machine, etc., the technical principles are similar.
In October, Hikvision released its third quarter performance report for 2019. During the reporting period, the company achieved operating income of 15.916 billion yuan, a year-on-year increase of 23.12%; net profit attributable to shareholders of listed companies was 3.811 billion yuan, year-on-year. The increase was 17.31%. The results were in line with expectations, and the expense ratio has also improved. Securities companies including Everbright have maintained a “Buy” rating on them.
Of course, it’s not all good news lately. In October, the United States added 28 Chinese companies and institutions, such as Hikvision, to the list of entities, and the company’s imports of US-originated goods, technology or software will be restricted. Hikvision subsequently stated that most of the US components can be directly replaced or replaced with new designs without affecting product performance or impact. For a small amount of material that is temporarily too late to replace, the company exchanges for inventory for a longer expected processing time. There are also a small number of materials that can be purchased and used by the customer by adjusting the business strategy.
The following is the conversation between the two media, including 36 与, and Xu Ximing. The content has been edited and edited without affecting the original intention.
Q: Many of the solutions that Haikang exhibited at AMB this year involved digital transformation of traditional industries. What role does Haikang play in this digital transformation? What are the technical reserves?
X: “Digital transformation” is not the same as “information transformation”, and informationization may be part of digitalization. Informatization relies more on the Internet and the mobile Internet, and digitization is about introducing the Internet of Things and artificial intelligence. Take a company as an example. The first step in the digital transformation of is to informatize, which means to informatize business processes and transaction data;
The second step is digitization, which refers to the digitization of equipment. From the application system, it is MES and industrial Internet of Things. In this process, Hikvision can provide “eyes” for digitization, including visual eyes, and other non-visual wavelength eyes, such as the Raytheon.
The third step is to digitize the scene and digitize the work, behavior, environment, people, cars, etc. In the scene digitization, Hikvision can provide a platform for enterprises. We hope to build a complete digital twin through the digitization of these three aspects to help users understand the rules in the middle.
We want to be the user on our platformThere are four things to be solved: the first to close the management distance, the second to improve management efficiency, the third to standardize the behavior, and the fourth to prevent security risks.
Q: Many companies at the AMB have installed algorithms on the chip to make hardware such as servers and smart access control. How do you think about the productization of many technology companies? Many people say that this industry is very simple, do you agree?
X: This is a very complex industry. Because we haven’t seen any company that has been driven by capital. If it is a very simple industry, the head company is very easy to wildfire under the driving of capital.
I chatted with a CIO of a head car company yesterday. He asked me why Hikvision always has two project managers. I said that because this business is different from traditional IT. Traditional IT projects can be done with just one project manager. But when we are doing digital transformation, there are traditional informationization work to be done, such as user communication, change management, project planning, research and development, as well as delivery, debugging, installation, installation angle, data sampling, etc. A lot of complicated work.
Some things have been done by many companies, and the face of the door was released the year before. But there are also cost issues, installation problems, and service issues. So the final competition of is to see who can really create value for users, who can have the best price/performance ratio.
The Internet of Things industry is harder than the IT industry, and I have a say. Throughout the process, you have to consider the factors and do more things. Moreover, fault tolerance is smaller.
For example, we help a company to do face-to-face access control attendance. In the morning, this access control machine is broken. There are hundreds of workers gathered at the door to get out of work. What is the concept? In the past, you did an attendance system, and it took two hours to get down. It won’t be a big deal. But the face is locked for five minutes, and the customer is coming to talk to you about how much it costs.
From IT to AIoT, delivery is much more difficult than before.
Q: The most talked about at the Ali Yunqi Conference is the urban brain. They are going to help the airport to do the dispatching, and also help the factory to reduce costs and increase efficiency. What Haikang does? What is the difference with them?
X: Our borders are limited, mainly for scenes. I divide the digitization of the enterprise into three parts: The first is informationization. In terms of traditional informationization, Ali has done a lot of things, especially in dealing with this part of the cloud. The second is the equipment association. This involves MES and the Industrial Internet of Things. The third is the scene of the object. Hikang will do a few things in the scene of the object: First, the Internet of ThingsSensor connection; second, the perception of the sensor of the Internet of Things, ie AIOT; third, on this basis, data aggregation, ie wireless convergence. Fourth, data insights, we provide some of our own technology and products.
Q: In the future, Haikang will continue to spend more effort on digital bases and cloudization, or will it work with other vendors to do things?
X: The user’s architecture and system selection, the decision is in the user, not in the vendor. Vendors can only provide solutions based on their own strengths and capabilities, allowing users to choose. The user has to choose the digital base, and the user has the final say.
The concept of a digital base is proposed by the manufacturer. Users can choose or not. Just like the concept of today’s scenes, I mentioned it, users can choose or not. What we value more is the value that we do to the user. Is it that I can help it to close the management distance, can it help it improve management efficiency, can help him regulate the operation and behavior, which is our concern. We are still adjusting to the needs of users.
Q: The AI camera launched by Huawei and other manufacturers at the AMB meeting felt that there were overlaps with some of Haikang’s products. Haikang’s main products, such as the Lei Shi machine, are not difficult to copy? Is the technical threshold high?
X: The difficulty of imitation is not high, it is not what we said. I feel that the company still needs constant innovation. As long as a company has continuous innovation, it is not afraid of competition. Haikang has a very good gene, which is to encourage innovation. As long as there is continuous and continuous innovation, even if we don’t make cameras after ten years, the company will still live very well.
Q: When it comes to encouraging innovation, the third-quarter earnings report shows that Haikang’s R&D expenses are growing faster than revenue. How does this affect your business group? Will sales pressure be great?
X: The increase in R&D expenses is to match the growth of the company’s sales. In the past two years, Haikang has implemented the “business sinking” strategy. We have turned provincial branches into provincial-level business centers, pushed forward resources, and shifted the focus of business decisions from headquarters to provincial-level business centers. During the migration process, the R&D and marketing personnel of all provincial business centers are growing, and the resources are sinking to the secondary city branch.
After experiencing a large investment cycle, we will enter a relatively stable personnel investment cycle.
In terms of sales pressure, we have always had a goal, whether it is good or bad. From our perspective, we still like to have a solid sales performance.
Q: There are a lot of AIoT companies talking about open platforms at AMB. What are the characteristics of Hikvision’s AI open platform?
X: We have three main points. First, our algorithms are straightforward and can be deployed directly to a camera or to a hard disk recorder, unlike many open platforms from many vendors. Second, as long as it is Haikang’s equipment, there is no charge for training and deployment on the platform. We hope to make AI enjoyable. Third, we provide components to help partners get through the application.
Q: From the perspective of your business unit, will Haikang’s investment in hardware and software be changed in 2020?
X: The changes in the company’s specific investment depend on market changes. The key to promoting the industry’s application is to empower application partners. For example, we have developed a component in the unified architecture to help users manage the model and version of the training on the AI open platform, and then combine the components and application software to record the false positives and false negatives. . This approach empowers all walks of life and helps partners have more capabilities.
Q: We found that the speed and quality of AI’s intelligent landing did not seem as good as expected. From the 2015 AI to more than four years this year, the camera should be the best landing, security industry is also the best landing, but the current intelligent camera in this industry is less than 5%, this speed It can be said that it is not ideal. How do you feel about this problem?
X: That’s why we are launching the AI open platform. Because many scenarios are fragmented, we hope to empower the entire industry with the AI open platform. At present, there are many partners in the Haikang AI open platform, and the total number has reached one thousand.
Q: Do you think there will be a breakthrough in the future?
X: Of course, this is an important engine for our EBG performance growth this year. AIOT must be a business that is going to be solid.
Q: That means you still have a lot of solid grounds?
X: For example, the factory’s morning face access control attendance system, maybe A factory can, B factory is not necessarily OK. A factory may have a shed, the lighting conditions are very good, butIt is a serious backlit of the B factory. The C factory is very crowded in the morning. The D factory may be in the research institute. It is a high-security zone. The F factory has foreign employees. The N factory is a coal mine. The faces of the brothers from the well are black. of. How do you get these problems?
Some fragmentation must be solved through services, applications, and system design, but some fragmentation is to be solved by ubiquitous technologies like the AI open platform. This is definitely not a single point, not just an algorithm. Equipment, services, installation, and applications must all be combined.
Q: Before Haikang said that if necessary, he would make the chip himself. Has this work been carried out now? Which part of the first product might be the chip?
X: The specifics depend on the disclosures that follow. Thank you.
Source of the title map: Hikvision