Unicorns in the field of AIoT are experiencing a wave of listing

Recently, there has been a wave of unicorns in the field of AIoT that are about to go public or are about to be "listed". Representative companies that took the lead include Megvii Technology and Cambricon, and CloudWalk, SenseTime, and Unisound have also "been" "listed" one after another. The outside world has always maintained great curiosity about the real development of these AIoT unicorns, and in this article we will analyze them one by one.

Recently, unicorns in the AIoT field have ushered in a wave of companies that are about to go public or are about to be "forced" to go public.

Representative companies taking the lead include Megvii Technology and Cambrian, while Yitu, SenseTime, and Unisound have also "been" forced to disclose their listing schedules.

Why are these companies in the AIoT smart Internet track?

We can get some clues from how they define themselves.

This week, AI chip company Cambrian's application for listing on the Science and Technology Innovation Board was approved.

Cambrian's prospectus did not mention the Internet of Things or AIoT in the first place, but founder Chen Tianshi mentioned in an interview:

"In the future, the Internet of Things and AI will complement each other and be inseparable. The Internet of Things brings life to all kinds of devices and terminals, while AI gives them intelligence... The upcoming 5G era will further accelerate the explosive progress of technologies such as the Internet of Things and AI. The physical objects connected by the Internet of Things are diverse and the application scenarios are rich... All these intelligent data processing in all scenarios requires the participation of AI chips."

In its prospectus, the Internet of Things is presented in a different way:

"The product system launched by Cambrian covers the cloud, edge and terminal... radiating to "smart +" industries such as smart Internet, smart manufacturing, smart transportation, smart education, smart finance, smart home, and smart medical care."

The founder of Megvii Technology mentioned in the opening of the prospectus:

"In the past eight years, we have researched and launched many artificial intelligence applications with practical value, including facial recognition technology, and we are more determined to believe that the Internet of Things is the main application scenario of artificial intelligence technology..."

“Our vision is to build an AI infrastructure that connects and enables tens of billions of IoT devices.”

With the disclosure of the prospectus, the financial, business and other conditions of these unicorns have clearly surfaced.

In the AIoT era, the process of artificial intelligence from technology iteration to application implementation is gradually accelerating.

Technological innovation is no longer the only criterion for measuring AI. How to gain market recognition and create industrial value has become the journey that AIoT unicorns are heading for.

The outside world has always maintained great curiosity about the real development of these AIoT unicorns. In this article, we will analyze:

What is AIoT? Why have so many startups emerged in the AIoT field in recent years?

AIoT unicorns have generally completed multiple rounds of financing, raising at least hundreds of millions of dollars. What core competitiveness have they invested in? What is the difference between them and traditional software and hardware companies? Compared with "traditional enterprises", what advantages do AIoT unicorns have in terms of financial data?

Do AIoT unicorns have disruptive innovation in their business models? In addition to technological innovation, what is the real value of AIoT for all walks of life?

Why are there so few news about AIoT unicorns going public abroad?

01

What exactly is AIoT?

With the continuous development of science and technology, some technologies that complement each other in function are naturally forming a "match made in heaven": for example, artificial intelligence AI and the Internet of Things IoT are one of the "good matches".

Due to the rapid development of the Internet of Things (IoT), a large number of devices or "things" in enterprises have achieved network connection and data sharing. Since AI can "learn" from massive amounts of IoT data to make quick decisions and build deep insights, it is an indispensable analytical capability for enterprises that want to expand the value of the Internet of Things (IoT).

In the book "Intelligent Internet of Things: New Thinking", I gave the definition of the Intelligent Internet of Things (AIoT):

AIoT is built on the foundations of the Internet, big data, artificial intelligence, and the Internet of Things. It is an intelligent Internet that connects everything and is an important carrier and way of thinking in the intelligent era. AIoT abstracts the physical world into the model world, and uses it to establish a complete digital world and build a new type of production relationship. AIoT changes the old thinking mode, thereby realizing large-scale social collaboration between people, people and things, and things and things.

Market research firm Markets and Markets recently released a report stating that the global AIoT market size was $5.1 billion in 2019. By 2024, this figure will grow to $16.2 billion, with a compound annual growth rate of 26%. Markets and Markets also pointed out that the need to effectively process the large amount of real-time data generated by IoT devices is the main driving force for the growth of the global AIoT market.

According to iResearch's research report "Artificial Intelligence 2020: Implementation Challenges and Responses", the artificial intelligence industry can be divided into three parts: basic layer, general layer and application layer, according to the logic from underlying basic technology to upper-level industry applications.

The basic layer provides computing power support such as chips and computing frameworks for basic artificial intelligence technologies such as image and voice. The general layer provides general technologies such as perception and cognitive computing. The application layer is products and services that generate application value by deeply integrating general artificial intelligence technologies with various industries.

Basic layer: It provides computing power support for basic artificial intelligence technologies such as computer vision and speech recognition. It is the infrastructure of artificial intelligence, including AI chips, AI platforms, and AI computing frameworks.

Typical representative companies include Cambrian, Huawei HiSilicon, Suiyuan Technology, Horizon Robotics, etc.

Artificial intelligence technology is widely used in cloud, edge, and end devices in the field of AIoT, and all of them require computing power support from core chips. The three scenarios of cloud, edge, and end have different requirements for chip computing power and power consumption, and a single category of smart chips is difficult to meet practical applications.

Taking Cambrian as an example, it has developed three types of chip products for the three scenarios of cloud, edge and end, namely terminal intelligent processor IP, cloud-based intelligent chips and acceleration cards, and edge intelligent chips and acceleration cards, and has developed a unified basic system software platform for all products in these three product lines.

Currently, "AI chips" have become one of the most watched tracks in China's innovation market. China has more than 30 newly established AI chip design companies, which have attracted a large amount of funds in the primary market.

General layer: General artificial intelligence technologies and products developed based on infrastructure, such as computer vision algorithms and robotic systems, are mainly divided into two parts: general software technologies represented by perception computing and cognitive computing technologies, and general products that integrate hardware and software, such as drones and robots.

Thanks to the rapid development of deep learning technology, in the past few years, general artificial intelligence technologies such as speech recognition and machine vision have "surpassed" human levels and quickly entered the historical stage.

In 2015, Microsoft achieved a milestone when it developed a computer vision system that could identify objects in images more effectively (95.1% accuracy) than humans (94.9% accuracy).

Therefore, a number of AIoT start-ups have emerged around voice recognition and machine vision.

Typical representatives in the field of speech recognition include: iFlytek, AISpeech, Unisound, and Mobvoi.

Typical representatives in the field of machine vision include the "Four AI Dragons": Megvii, SenseTime, CloudWalk and Yitu.

Megvii Technology has its own deep learning framework and mainly provides personal IoT solutions, urban IoT solutions, and supply chain IoT solutions.

SenseTime has an original underlying algorithm platform and relies on the platform to empower other industries with technology.

Yitu is a national enterprise, focusing on banking, finance, security, and transportation; Yitu is deeply engaged in the medical field and gradually expanding into security and finance.

ArcSoft focuses on the smartphone field and is expanding into smart cars and smart homes.

Hikvision and Dahua are traditional security giants, both of which focus on security business and are expanding into new areas.

Application layer: The field where general AI technologies are deeply integrated with applications in various industries, mainly vertical AI application companies. Application layer AI companies package general technologies into products that can be implemented, including integrated hardware and software products and end-to-end solutions for specific application scenarios.

As general technologies mature and their industry application value becomes more prominent, a large number of general-layer companies are also extending to the application layers of various industries based on their basic technical capabilities.