LG Electronics Production Engineering Research Institute (LG PRI) has reportedly received its first order for a laser direct imaging (LDI) tool for semiconductor packaging. According to South Korea's Electronic Times, the order came from a global OSAT (outsourced semiconductor assembly and test) company that also operates packaging plants in Korea, and the tool is intended for a production line. For LG PRI, which had previously supplied the equipment for university R&D use, this marks the first purchase order (PO) obtained for an external customer's production equipment.
However, what Electronic Times reported extends only as far as the PO. Neither LG PRI nor the customer has officially announced the order, and the customer's name and the number of units remain undisclosed. Price, delivery schedule, and installation site are also unpublished. A PO is not delivery, acceptance, process qualification, or mass-production operation. This development should be read not as a story about LG manufacturing semiconductors, but as LG becoming a candidate vendor supplying equipment to an OSAT's production line.
From R&D Supply to a Production-Line Order
According to Electronic Times, LG PRI introduced its semiconductor packaging LDI system in 2025. It initially supplied the tool for university R&D applications, and this latest order is reportedly the first PO for an OSAT production line. For manufacturing equipment, there is a large gap between being used in R&D and entering a customer's mass-production process. The latter requires installing and accepting the equipment according to packaging materials and product-specific conditions, qualifying the process, and running it stably in mass production.
At this point, none of those later stages have been reported. It is also unclear which LG PRI model was covered by the first PO. Electronic Times reported that LG PRI also developed 3µm and 5µm systems to meet customer requirements, but it has not disclosed which model was ordered this time. Therefore, it cannot be said that a 1.5µm-spec tool was ordered, or that its performance has been proven in mass production.
The change reported here is narrow but clear: following its supply of research equipment to universities, LG PRI's LDI has advanced to the stage of being ordered as production equipment by an external customer. Actual equipment sales and mass-production track record can only be assessed once delivery and acceptance occur.
Exposing Resist, Not Drawing Metal
LDI is an exposure tool that projects light onto photoresist based on digital patterns, without using a physical photomask. The resist that remains after development defines where metal wiring will be formed in later process steps. The LDI tool itself does not directly draw metal.
LG PRI's official page lists a line/space (L/S) of 1.5µm as the top published specification for this system. The company states that it generates and corrects patterns in real time, combining projection lenses, precision alignment, and stage technology. Applications include advanced semiconductor packaging as well as displays and MEMS. Published use cases include PDP, LCD, and OLED displays, along with R&D PCBs.
Maskless exposure reduces the need to remake physical masks each time a design changes, allowing patterns to be corrected digitally. However, LG PRI has not disclosed throughput, mask-cost savings, process-time reduction, or yield-improvement figures. How effective the tool is in mass production cannot be judged from resolution numbers alone.
1.5µm Is Fine, But Not the Number That Determines Production Performance
Comparing published specifications, LG PRI's 1.5µm L/S is finer than the 2µm line width/2µm spacing that TSMC has disclosed for CoWoS-R. Meanwhile, SCREEN's DW-3100 claims sub-1µm accuracy and offers three types of alignment—global, local, and die-by-die—along with wafer or rectangular panel configurations.
| Equipment/Technology | Published Miniaturization Metric | Published Conditions |
|---|---|---|
| LG PRI Semiconductor Packaging LDI | L/S 1.5µm | Real-time generation/correction, projection lens, precision alignment, stage technology |
| TSMC CoWoS-R | Minimum pitch 4µm (2µm line width/2µm spacing) | In mass production since 2023 |
| SCREEN DW-3100 | Sub-1µm accuracy | Three types of alignment, wafer or rectangular panel configuration |
This comparison alone cannot be used to conclude adoption in CoWoS-R or process compatibility. LG's 1.5µm L/S is neither an overlay accuracy figure nor a minimum pitch figure, and conditions such as materials, resist thickness, and aspect ratio are not aligned. Substrate sizes also differ. Furthermore, mass production demands scrutiny of throughput and defect density. Overlay accuracy, yield, and process qualification also need to be confirmed. LG's published specification falls into the fine-pitch category, but there is no basis for claiming it represents the highest resolution currently available in the market.
The fact that SCREEN presents three types of alignment separately from its miniaturization metric illustrates well why mass-production equipment cannot be selected on resolution alone. The method for correcting positional misalignment differs across the whole panel, local regions, and individual dies. LG PRI also claims precision alignment technology, but has not disclosed its method or accuracy. Being able to draw a 1.5µm line is a different capability from how precisely that line can be overlaid onto the wiring or connection terminals of the preceding layer.
CoWoS is also not a single wiring structure. CoWoS-R uses an RDL interposer, while CoWoS-S adopts a silicon interposer. CoWoS-L combines RDL with high-density local silicon interconnects. Even if LG's published L/S is close to the wiring dimensions of CoWoS-R, this does not mean it can handle the processes for forming CoWoS-S's silicon wiring or CoWoS-L's local silicon interconnects.
Not Yet a Direct Solution to the CoWoS Shortage
TSMC itself has acknowledged that advanced packaging capacity is tight. At the company's Q2 2026 earnings call on July 16, CEO C.C. Wei explained that packaging capacity shortages are limiting customer growth. He said the company is working to narrow the gap between demand and capacity, and that back-end processes remain in shortage with a significant supply-demand gap.
However, this customer is an anonymous OSAT, and Electronic Times did not report any connection to TSMC, CoWoS, CoWoS-R, or any specific AI chip. TSMC's 2026 capital expenditure is projected at $60 billion to $64 billion, of which 10% to 20% is allocated to a combined category covering advanced packaging, testing, mask-making, and other items. The company has also explained that bottlenecks can shift between front-end and back-end processes, and that tester shortages can also occur.
LDI handles just one step—exposure—within the back-end process. A single PO from an anonymous OSAT is neither evidence that TSMC's CoWoS supply capacity is increasing, nor a solution to the advanced packaging shortage. What should be confirmed first is which process, and with what configuration, the equipment will be delivered for.
LG PRI's Exposure Technology and What to Confirm Next
LG PRI was founded in 1987 and has developed a laser patterning system for PDPs in 2002, an LCD color filter exposure system in 2006, and a smart exposure system for displays in 2014. In 2021, it added a semiconductor packaging inspection system. The semiconductor packaging LDI represents less a from-scratch entry into exposure equipment than an extension of this existing exposure and inspection technology into external semiconductor applications.
Electronic Times reported that LG PRI is emphasizing price competitiveness and plans to expand into HBM inspection equipment and TGV laser equipment for glass substrates going forward. However, pricing and launch timing for each plan remain undisclosed. Customers and production status are also unclear, and these plans should not be conflated with the current LDI's track record.
What has been confirmed from this PO is only that it was reported as an order for a production line. The next indicators to watch are whether delivery and acceptance actually occur, and how much of the mass-production metrics—throughput, overlay accuracy, defect density, and yield—get disclosed. Only once those become visible can we gauge whether LG PRI's LDI has the potential to establish itself as production equipment in the back-end process, moving beyond its origins as an R&D tool.