Last year AI was a negative label. The combined loss of SenseTime and other AI tigers exceeds 50 billion yuan, which frightens many investors. This year, with the launch of ChatGPT, AI companies bottomed out. A boss told the reporter of Southern Finance All Media that financing was tight last year, but this year investors are lining up to give money, and in the commercialization of the landing, there are a lot of definite direction.
At present, in the commercial application of AI, the most mature is machine vision, which acts as “eyes”. Our brains are wired to get 75% of our information from sight. In reality, the combination of machine vision and industrial scene is higher. Recently, the popularity of AI in industrial quality control applications is increasing.
At present, industrial AI quality inspection can solve the two major pain points in the production and manufacturing of 3C consumer electronics, automobiles and parts. One is the difficulty in recruiting quality inspection workers. Usually, quality inspection workers’ eyes should be directly exposed to strong light for more than 10 hours a day, which is easy to cause workers’ eyesight to decline rapidly; The other is that manual inspection cannot guarantee stable output day after day. If it encounters tens of thousands of semiconductor wafers at every turn, it will bring great loss to the factory.
The above pain points let a number of companies see a blue sea to explore. In the panel industry, 80 percent of quality inspectors can be replaced by AI, Jiang Yuan, senior architect of Gechuang Dongzhi AI, said in an interview. Huang Qiang, product director of Tencent Cloud Industry AI, said that Tencent sees a 100-billion-level rigid demand market in AI quality inspection. “There are nearly 3 million people involved in the manual testing of 3C electronic products alone. If the annual cost of employing a person is 80,000 yuan, it will be a market of 240 billion yuan.”
According to IDC, the overall market of industrial AI quality inspection in China will reach $958 million (about 6.2 billion yuan) by 2025, with a CAGR of 28.5% from 2021 to 2025. At present, 100 billion market is far away, industrial AI quality inspection market is still several times the space, which makes more companies eager to try, accelerate into the battlefield.
In this year’s increasingly fierce market competition, IDC believes that industrial AI quality inspection market will continue to be fragmented. Currently, the market is dominated by cloud service providers, AI quality inspection innovation enterprises and machine vision software enterprises. And industrial Internet platform service providers, machine vision inspection equipment manufacturers, integrated AI enterprises, industrial control automation enterprises, communication operators are increasing investment to enter the market. An “inner coil” between machines is unfolding at an accelerating rate.
Startup: Taking advantage of Silicon Valley’s new technology, a former Microsoft executive “reduces the dimension” to enter the AI industry quality inspection
Guo Wei, former senior director of Microsoft’s SurfaceBook design team, has contributed more than 30 invention patents for Microsoft. In 2018, he immediately gathered a group of overseas elites who are good at data processing, platform technology and AI algorithm to start their own business in China after seeing the trend of AI scenario-based application in China.
It is worth mentioning that in the current wave of AI commercialization, the market is divided into three levels: the bottom infrastructure to make chips and computing power, the middle core technology to make large models and big data platforms, and the upper scenario-based application to make solutions landing. In addition to technology, chip also emphasizes ecology, and computing power is a competition between countries. Big models and big data are in the hands of giants who can afford to burn money. After analyzing and excluding one by one, the most advantageous place for startups in China is to make application.
Guo Wei founded Bohan Intelligence in late 2019 in Nanshan district, Shenzhen. Based on his resume, the team he formed won the pre-A round of financing of 35 million yuan from Shenzhen high-level talent venture funding and investment institutions, and jointly established the artificial intelligence Research and development Center with Shenzhen Tsinghua Research Institute.
He positioned the company as a company that relies on self-developed platform technology and focuses on the implementation of AI applications, focusing on intelligent manufacturing and autonomous driving. In the benchmarking field, intelligent manufacturing works closely with Huawei (NPU ecosystem), and autonomous driving is integrated into Nvidia GPU ecosystem. “The company’s technology and products are honored to be recognized by the strongest partners in the industry and have deep business cooperation with them”
Backed by Huawei computing base, Guo Wei’s team is determined to take AI quality inspection in the manufacturing field to a new level, namely to complete the “impossible triangle” in Party A’s mind — shorter development cycle, higher accuracy, and lower operation and maintenance cost, so as to subvert traditional AI quality inspection.