GUMI, South Korea – On a recent visit to LG Innotek's "Dream Factory" in Gumi, North Gyeongsang Province, the state-of-the-art facility dedicated to the production of Flip-Chip Ball Grid Array (FC-BGA) substrates, the initial encounter within the sterile environment was not with human workers, but with a fleet of autonomous mobile robots (AMRs). After donning protective suits and undergoing an air shower, the scene inside the plant revealed dozens of these robots seamlessly navigating the rows of sophisticated equipment, autonomously transporting various raw materials to the processing stations. The robots operated based on pre-programmed specifications and delivery schedules inputted by customers.
The few human personnel present primarily oversaw the smooth operation of these robotic colleagues. The intricate ten-step manufacturing process was largely executed by these tireless machines. Even the critical final stage – quality assurance – was entrusted to robotics. A gripper-equipped robot precisely moved the completed golden-hued substrate to an inspection device. Within 30 seconds, an integrated artificial intelligence (AI) deep learning vision system meticulously analyzed the product, determining its conformity to quality standards.
Historically focused on manufacturing camera modules for smartphones, LG Innotek has boldly positioned the FC-BGA business as its next-generation core growth engine. The company's acquisition of the Gumi plant from LG Electronics in 2019 was followed by a groundbreaking move: the introduction of an industry-first, cutting-edge AI-driven automated production system. This strategic investment amounts to a staggering 1.4 trillion won (approximately $1 billion USD). As a relatively late entrant to the FC-BGA market in 2022, LG Innotek recognized the imperative of possessing a decisive competitive edge to effectively challenge established players. The company's strategy hinges on leveraging differentiated product competitiveness to ascend to the ranks of global top-tier FC-BGA manufacturers.
Lean Workforce Powered by AI Automation
The cornerstone of LG Innotek's Dream Factory is the minimization of human intervention and the pervasive integration of AI-powered automation. Recognizing the extreme sensitivity of the FC-BGA manufacturing process – where even microscopic scratches or human respiration can significantly impact defect rates – LG Innotek has fundamentally restricted human access to the production line. The entire process is instead managed by a sophisticated ecosystem of AI, robots, and deep learning technologies and equipment.
FC-BGA substrates are advanced semiconductor packaging solutions primarily utilized in high-performance computing applications such as Central Processing Units (CPUs), Graphics Processing Units (GPUs), and communication chipsets. The burgeoning demand for these substrates is fueled by the rapid growth in AI, autonomous driving, and extended reality (XR) technologies. However, the manufacturing process is notoriously complex. Product design and process intricacies result in significantly lower yields compared to conventional printed circuit boards (PCBs), with typical FC-BGA yields averaging around 50%, far below the 95% seen in standard PCBs. LG Innotek is betting on its ability to maximize yield rates through technological prowess, thereby establishing a strong competitive advantage despite being a latecomer to the market.
A key element in achieving this enhanced yield is the AI vision inspection system. This critical final checkpoint automatically verifies whether the manufactured product precisely meets the customer's specified requirements. Known as Line Quality Control (LQC), this system can identify defective products and autonomously sort them without human intervention. Crucially, the inspection data is immediately transmitted to the customer, ensuring unparalleled quality transparency. Park Jun-soo, head of LG Innotek's FS Production Team, explained, "This process has enabled us to reduce the lead time from order to delivery by up to 90%, and we've also been able to decrease the personnel involved in sampling inspections by 90%." Furthermore, AI is integrated into a digital simulation system that proactively predicts potential product defects and equipment malfunctions, significantly reducing the time previously spent by human operators manually checking for anomalies.
The adoption of a "digital twin" system further contributes to yield optimization. Before physical deployment, all production equipment undergoes virtual factory simulations using 3D modeling. This allows LG Innotek to identify and rectify potential issues with the production line in a virtual environment, ensuring the deployment of an optimized and efficient physical setup. As a result, LG Innotek has been able to achieve high initial mass production yields and significantly shorten the ramp-up period for production expansion by nearly half compared to traditional methods.
Kang Min-seok, Senior Vice President and head of LG Innotek's Substrate & Material Business Division, emphasized, "Dream Factory differentiates itself from competitors through its AI-driven foundation, integrating all of LG Group's technological capabilities. We have not only achieved a significant increase in yield but also operate with approximately 50% fewer production personnel compared to our competitors."
Looking ahead, LG Innotek aims for complete automation of its entire FC-BGA production process, striving for zero human intervention, including not just manual laborers but also engineers. To achieve this ambitious goal, the company plans to implement a process intelligence system by 2026 that will enable real-time detection and analysis of quality deviations, automatically implementing corrective actions.
Targeting Trillion-Won Business by 2030 with Further Investments Planned
Industry observers are keenly watching LG Innotek's potential to gain a significant foothold in the FC-BGA market, currently dominated by Japanese giants Ibiden and Shinko Electric Industries. The competitive landscape also includes major players like Taiwan's Unimicron, Austria's AT&S, and South Korea's Samsung Electro-Mechanics. According to Fuji Chimera Research Institute, the global FC-BGA market is projected to grow from $8 billion USD in 2022 to $16.4 billion USD by 2030, indicating substantial growth potential.
LG Innotek has set an ambitious target of scaling its FC-BGA business to a trillion-won (approximately $720 million USD) scale by 2030. Following its initial supply of FC-BGA for PCs to a global Big Tech company at the end of last year, the company recently secured additional major global clients. Its goal for this year is to enter the PC CPU FC-BGA market, with the break-even point anticipated in the following two years. LG Innotek intends to leverage its extensive global network, built over nearly three decades in the substrate materials business, to achieve its objectives.
Further investments are under consideration. Following a recent announcement of a 600 billion won (approximately $430 million USD) additional investment to expand its business, LG Innotek plans to further increase its production capacity based on securing additional customer orders. Including the initial plant acquisition, the total investment is expected to surpass 2 trillion won (approximately $1.44 billion USD). LG Innotek is also actively collaborating with global Big Tech companies with the aim of commercializing next-generation glass substrate technology by 2027. This foray into advanced materials further underscores LG Innotek's commitment to innovation and its ambition to become a leading force in the high-performance substrate market. The company's aggressive investment strategy and focus on cutting-edge automation position it as a formidable challenger to the established leaders in the rapidly expanding FC-BGA sector.
[Copyright (c) Global Economic Times. All Rights Reserved.]
