Author | Li Wei
IT people often look at the semiconductor industry and often feel incredible - for example, Tencent will not be inseparable from Ma Huateng, Alibaba will not be inseparable from Ma Yun, but the research and development of advanced processes is inseparable from the research and development of generals with deep industrial experience.
Before engaging in semiconductor intelligent software, Li Jie, CEO of AX Industries, listened to friends from time to time to share his semiconductor industry insights. He learned that the semiconductor industry's high dependence on people is a major pain point. "Can't the experience of these big guys form a system?" If ai and other technical means are used to liberate knowledge workers, it will be a huge market opportunity. ”
Li Jie studied automation and industrial engineering as an undergraduate, and obtained a degree in computer science during his master's degree, specializing in big data, and later read a doctorate in intelligent science and systems engineering, and has been thinking about how to cross the disciplines he has learned and solve problems in some subdivisions. Coupled with the doctoral stage, he had followed his mentor to participate in semiconductor intelligent manufacturing projects, and used engineering theory to solve semiconductor modeling and scheduling problems, which also gave him the idea of starting a business around the semiconductor field.
In 2017, Li Jie co-founded AX Industries and officially stepped into the semiconductor industry. AX is positioned as a semiconductor intelligent manufacturing solution provider, and the original intention of using technical means to solve industry problems is written in the company's name. The four letters of "IKAS" represent industry, knowledge, automation and system, which means to use intelligent manufacturing systems carrying knowledge to help the industry achieve automation and intelligent upgrading.
AX mainly provides industrial software for semiconductor and pan-semiconductor manufacturing industries such as integrated circuit manufacturing, packaging and testing, LED, photovoltaics, etc., and reduces costs and improves efficiency for scheduling and equipment maintenance in the manufacturing process. According to reports, AX currently serves 80 to 90 customers at the same time, and semiconductor-related business accounts for 70% and focuses on wafer foundry customers.
The value of cost reduction and efficiency improvement should be written into the contract
The core team of AX Industries presents the characteristics of Trinity, which is mainly divided into a team of founding scientists, a team of industry experts composed of senior experts at home and abroad, and a team of AI and big data research and development. Overall, self-developed ROPN industrial system modeling and artificial intelligence are the company's two underlying technologies.
Based on these two core technologies, AX Industries has built three major intelligent manufacturing product lines of intelligent equipment, intelligent production line and yield improvement, covering nearly 50 software products and functional modules as a whole. The company's software products penetrate into the entire semiconductor production process, covering multiple core manufacturing areas such as asset monitoring and control, yield improvement, and intelligent scheduling, and empower manufacturing from core KPIs such as asset utilization, production time, production capacity, and cost.
Li Jie introduced that with the same use of big data technology, compared with the well-known artificial intelligence startups in China, AX Industry is based on specific market segments to solve problems, and emphasizes what value is realized for customer business. Semiconductor intelligent manufacturing software focuses on automation and intelligent upgrading, in addition to domestic substitution, it actually focuses more on the vast incremental market of intelligent manufacturing.
Li Jie specifically mentioned that in recent years, under the background of the continuous shortage of semiconductor production capacity, customers pay special attention to products that help increase production capacity. Increasing existing production capacity through intelligent production lines is the urgent need of customers that the company sees.
Using the APS Smart Scheduling Software System to boost fab output is a classic use case for the company's products. According to reports, the bottleneck equipment group of a 12-inch fab in China faces pain points such as different equipment road widths and inconsistent output rates (WPH) of the same formula on different available equipment. AX replaced the RTD dispatch rules used by enterprises for many years with self-developed APS algorithms, which not only optimized the above problems, but also increased the total output by 2 to 3% under the premise of 100% equipment utilization rate.
In addition to the structural tight supply and demand, the shortage of talents is already the biggest challenge facing the semiconductor industry. The company observed the impact of the shortage of talents in core positions such as equipment, technology, yield, etc., and cooperated with industrial customers to liberate manpower with software systems, which not only reduced manpower by more than 90% in some scenarios, but also avoided the yield and scrap problems caused by human error. At present, several core systems have been adopted by nearly ten companies, including large wafer foundries and compound semiconductor manufacturers.
Li Jie introduced that how much production capacity or yield has increased is the indicator that customers are most concerned about, and the company will also write these data directly into the contract.
Optimistic about the tens of billions of scale equipment intelligent market
In Li Jie's view, the semiconductor intelligent manufacturing industry is an invisible market segment in the process of semiconductor manufacturing heavy assets. Although this market is currently not loud in the field of semiconductors, the market space is very broad. He also said that in the company's current three major product lines of intelligent equipment, intelligent production line and yield improvement, he is particularly optimistic about the products of the intelligent production line.
From the effect point of view, the machine as the basic production element, the asset efficiency improvement brought about by the intelligent equipment can be directly transformed into the improvement of enterprise productivity. In terms of the market, the stock equipment market alone has a large number of upgrade demands, and the overall scale is in the tens of billions.
In addition to intelligent equipment, AX Industries can see some forward-looking opportunities. Li Jie believes that China's IC industry, as a cutting-edge manufacturing industry, has always been closely linked to the IT industry, and its industrial upgrading is also rigorous and radical at the forefront. The rigorous side is that IC production is non-stop 24 hours a day for 365 days, and it will be very cautious about the introduction of new technologies. The radical side is that once there is a new IT technology, it is often the first time to be adopted by the manufacturer of the IC.
At the same time, industry leaders tend to have higher acceptance of intelligent manufacturing upgrades, which also makes AX's current customers mainly leaders such as SMIC and Northern Huachuang, so they can see cutting-edge market dynamics. The early layout of Industry 5.0 and the expansion of green transformation products in line with changes in the international policy environment are typical cases.
Li Jie revealed that the company's next focus is to use a highly generalized and standardized product system with Chinese characteristics to meet the different needs of users in the industry and empower manufacturing more efficiently.
He pointed out that standardization is actually something that international players such as Applied Materials and Siemens are doing. The smaller the degree of customization, the higher the gross profit of the product. He also said that AX will become an internationally renowned and domestic first-class semiconductor intelligent manufacturing solution provider in the future. This also means that product standardization is the only way to go. "A lot of software companies are essentially outsourcing companies for a niche. Without algorithms, models, software can not be SaaS (software as a service), not to help customers achieve the effect of the goal to position, then your company and software outsourcing company is no different. ”
He also stressed that making generalized products is only one of the means to enhance the company's competitiveness. At this stage, the company has 70% to 80% of the system and the underlying algorithm can be generalized, and there will be about 20% of the import and adaptation work for customers in each industry. "Compared with competitors whose delivery model is very solidified and cannot be flexibly adjusted according to user needs, 20% of the import adaptation services are our core competitiveness, and 20% of the services can make our product delivery effect and import cycle surpass that of international first-line manufacturers." He added that in fact, customization also involves customized efficiency propositions, similar to Haidilao's model, personalized services can also tend to be standardized.
epilogue
In recent years, the importance of the semiconductor industry in China is self-evident, but compared with large capital expenditures such as factory construction, intelligent industrial software is a growth area that was easily overlooked in the past. With the continuous strengthening of chip production capacity demand, local manufacturers' emphasis on production automation and intelligence has gradually increased, bringing opportunities to domestic semiconductor intelligent manufacturing solution providers such as AX Industries. Li Jie said that overall, chip manufacturing faces challenges such as tight delivery time, many production steps, complex technology, and large amount of production test data, which are all opportunities to provide products and services.