Printing a semiconductor chip sounds like science fiction. It is getting close to real. Seiko Epson and Taiwan's Manz Asia have joined forces to co-develop a next-generation process that manufactures semiconductors with inkjet. No photomask, the one thing conventional fabrication could not do without: the circuit material goes straight onto the substrate. A printing giant and a semiconductor equipment specialist make an unusual pair.
Two Companies, Two Strengths
On March 12, 2026, Seiko Epson and Manz Taiwan (known as Manz Asia) officially announced a strategic partnership to accelerate the adoption of inkjet technology in semiconductor manufacturing.
Epson is a household name in printers, but the real magic lies in its PrecisionCore printhead technology, refined over more than 40 years and manufactured using the same MEMS (Micro-Electro-Mechanical Systems) fabrication processes used to make semiconductor chips. Each nozzle can fire up to 50,000 ink droplets per second with precise volume control, using piezoelectric elements just one micrometer (0.001 mm) thick.
Manz Asia, headquartered in Taoyuan, Taiwan, specializes in equipment for semiconductor panel-level packaging (PLP). The company provides solutions spanning wet chemistry, plating, inkjet printing, automation, and software integration. It has gained particular prominence for its work in CoPoS (Chip-on-Panel-on-Substrate), a panelized evolution of the widely used CoWoS advanced packaging technology.
The partnership combines Epson's ultra-precise printhead technology with Manz Asia's deep expertise in semiconductor equipment engineering and process integration.
How Do You "Print" a Semiconductor?
To understand why this matters, it helps to know how chips are traditionally made.
The standard method is photolithography, a process where light is shone through a photomask (essentially a stencil of the circuit design) onto a light-sensitive material coating a silicon wafer. The exposed areas undergo chemical changes, and the unwanted material is then etched away to reveal the circuit pattern. Think of it as "subtractive" manufacturing: you start with a full layer and carve away everything you don't need.
Photolithography is extraordinarily precise, achieving features just a few nanometers wide at the cutting edge. But it's also extraordinarily expensive. Photomasks alone can cost hundreds of thousands of dollars each, the apply-then-remove cycle wastes significant materials, and the process consumes large quantities of chemicals and water. Lithography is estimated to account for roughly 30% of total IC manufacturing costs.
Inkjet semiconductor manufacturing flips this logic. It's "additive", meaning materials are deposited only where needed, building up layers like a 3D printer builds objects. Functional inks, conductive, photoresist, or specialty materials, are jetted directly onto silicon wafers, glass, or other substrates with no photomask required.
The advantages are compelling: rapid design iteration (no mask to remake), dramatic reduction in material waste, fewer process steps, and a significantly smaller environmental footprint.
Lab-to-Fab: Bridging Research and Mass Production
What makes this partnership particularly noteworthy is the "Lab-to-Fab" concept, creating a seamless path from laboratory research to factory-scale production.
In November 2024, Manz Asia established an inkjet lab at its R&D center in Taoyuan, Taiwan, equipped with Epson printheads. The facility has been fielding inquiries from semiconductor equipment manufacturers exploring inkjet-based process innovations, offering sample printing and technical consultation.
With this new agreement, both companies have committed to going beyond research and evaluation to build production-scale infrastructure. Manz Asia's equipment lineup tells the story: the "RDJet Series" handles R&D experimentation, while the "SDC Series" is designed for volume manufacturing. Together, they create a continuous pipeline covering R&D, pilot production, and mass manufacturing.
This means chipmakers can explore new applications, validate materials, optimize process parameters, and scale up to mass production without switching platforms or starting over.
Where Inkjet Shines: Applications and Possibilities
Inkjet technology is especially well-suited for advanced semiconductor packaging, the process of connecting, protecting, and enabling communication between chip components.
Manz Asia's inkjet solutions target several key applications: forming 2.5D and 3D antenna structures for RFIC (radio-frequency ICs), printing heatsinks for thermal management, and creating bonding layers, all for devices used in RFIC, PMIC (power management ICs), and CPO (co-packaged optics) applications.
The ability to directly deposit conductive ink patterns could replace portions of traditional plating and etching processes. And because the technology supports both 2D and 3D structure formation, it opens doors for the high-density, multi-layer packaging that AI chips increasingly demand.
Japan's Inkjet Edge: A Global Competitive Advantage
This partnership highlights a broader story: Japan's world-class inkjet technology is entering the semiconductor arena.
Epson's PrecisionCore printheads are precision-engineered components manufactured using semiconductor-grade MEMS processes. The company produces over 15 million printers annually and has unmatched experience in high-volume printhead manufacturing and quality control. In 2025, Epson also partnered with Germany's SUSS MicroTec to integrate PrecisionCore into semiconductor R&D printers, the Manz Asia deal extends that trajectory from lab to factory floor.
Japan's printed electronics ecosystem extends beyond Epson. Companies like Elephantech (which develops inkjet-printed circuit board technology) and Mimaki Engineering contribute to a robust domestic capability. The Japan Electronic Packaging and Circuits Association (JPCA) has co-authored international standards for printed electronics with IPC (the global electronics industry association), cementing Japan's leadership in setting the rules of this emerging field.
Will Inkjet Replace Photolithography?
Not entirely, and that's not the goal. Cutting-edge chips at 3nm or 2nm process nodes require nanometer-precision patterning that only EUV lithography and similar technologies can achieve. Inkjet resolution currently tops out at roughly 50 micrometers, orders of magnitude less precise.
Where inkjet excels is in the "back end" of semiconductor manufacturing, particularly packaging. Forming wiring traces, building insulation layers, creating thermal management structures, these processes need coverage over large areas rather than atomic-scale precision, making them ideal candidates for inkjet's cost advantages and process flexibility.
As AI drives explosive demand for advanced packaging, where multiple chiplets are connected in increasingly complex 3D arrangements, the case for additive manufacturing grows stronger.
The Bottom Line: From "Carving Away" to "Building Up"
The Epson–Manz Asia partnership represents a philosophical shift in semiconductor manufacturing: from subtractive to additive, from rigidity to flexibility, from waste-intensive to resource-efficient.
Japanese precision inkjet technology fused with Taiwanese semiconductor equipment expertise creates a platform that could meaningfully impact the global chip supply chain, not by replacing traditional lithography, but by offering a faster, greener, and more adaptable alternative for the packaging processes that increasingly define chip performance.
How is semiconductor manufacturing evolving in your country? What do you think about printing chips instead of etching them? We'd love to hear your perspective.
References
- https://corporate.epson/en/news/2026/260312.html
- https://www.prnewswire.com/news-releases/manz-asia-and-epson-form-strategic-partnership-to-advance-inkjet-technology-for-semiconductor-manufacturing-302711960.html
- https://eetimes.itmedia.co.jp/ee/articles/2603/16/news031.html
- https://www.manz.com.tw/en/about/
- https://corporate.epson/en/technology/overview/printer-inkjet/precision-core.html
Global Discussion
12 comments