Murata Manufacturing alone holds more than 18,000 active patent families worldwide, and a substantial cluster — the multilayer ceramic substrate filings that underpin modern semiconductor packaging — sit in commercialization windows aligned almost perfectly with Penang’s roughly $13B advanced-packaging build-out announced between 2024 and 2026. The substrate is the layer beneath the chip, and it is one of the least-discussed bottlenecks in ASEAN’s pivot from back-end assembly to true advanced packaging. Japan owns the substrate IP. Penang owns the floorspace, the labour, and the customer demand. The licensing pipeline between them barely exists.
Why substrates matter more than chips for Penang’s next phase
Penang’s semiconductor industry has lived off Outsourced Semiconductor Assembly and Test (OSAT) for forty years — wire-bonding, encapsulation, final testing, a labour-intensive but technologically modest set of operations. The 2024-2026 wave of investment is qualitatively different. Intel committed roughly $7B for advanced packaging in Kulim and Penang. Infineon committed about $5B for a third silicon-carbide site in Kulim. AMD opened a design centre in Penang. Texas Instruments expanded its assembly footprint. Lam Research and Applied Materials added equipment hubs. What these investments require is not more wire-bonders. They require substrate supply.
Multilayer ceramic substrates — Low Temperature Co-fired Ceramic (LTCC) and High Temperature Co-fired Ceramic (HTCC) being the two dominant families — sit beneath the die and route signals while dissipating heat. For 5G mmWave front-end modules, automotive radar, silicon-carbide power devices, and high-frequency RF, the substrate determines yield. Today the global supply is dominated by four Japanese firms — Murata, Kyocera, Ibiden, and Shinko Electric — alongside Taiwanese substrate houses. ASEAN OSATs buy substrates from these vendors. They do not make them.
The expiring Japanese substrate portfolio
A material share of Japan’s foundational LTCC and HTCC patents are reaching commercialisation maturity between 2026 and 2030. Murata’s JP4720193 family covered core multilayer ceramic capacitor and substrate co-firing methods; its priority filings have already exited the 20-year exclusivity window in some jurisdictions, and the family’s national-phase entries across ASEAN are now reaching their own expiry sequence. Kyocera’s HTCC ceramic-packaging cluster around JP2007091505 and its successor filings expires across 2026-2028 in major ASEAN jurisdictions where parallel filings were prosecuted. NGK Insulators in Nagoya holds the silicon-nitride substrate IP that is becoming critical for silicon carbide power devices — which is exactly the technology Infineon is bringing to Kulim.
These are not toy patents. They cover dielectric tape formulations, via-fill metallisation, co-firing temperature profiles, and stack-up architectures that determine whether a substrate ships at 95% yield or 60% yield. Reproducing them from a printed patent is not trivial — process know-how matters as much as the published claims — but the patent expiration removes the most expensive single blocker: licensing fees on the foundational claims and the threat of an injunction in an export market.
What ASEAN OSATs actually need
Inari Amertron, Globetronics, UTAC, Amkor’s Malaysian operations, Thai NEX, and several Vietnamese subcontractors have been quietly evaluating whether to vertical-integrate into substrate production. The economics are clear. Substrate cost for a typical 5G RF front-end module can exceed the silicon die itself. Importing finished substrates from Japan or Taiwan adds three to six weeks of lead time and eight to fifteen percent to the bill of materials. For the high-volume programs Intel and Infineon are bringing to Malaysia, that gap is meaningful at the corporate-margin line.
What blocks the OSATs is not capital. Malaysia’s NTIS and Singapore’s EDB have both signalled substrate-investment incentives. What blocks them is technology access, specifically:
- LTCC dielectric tape formulations (today controlled by Murata, Ferro, DuPont)
- Via-fill paste chemistry (Kyocera, DuPont, Tanaka Kikinzoku)
- Co-firing furnace recipes — partly patented, partly trade secret
- Substrate-die interface qualification, which remains a case-by-case engineering exercise
The expiring Japanese patents address the dielectric, via-fill, and co-firing layers. The interface qualification remains a Japanese engineering-services opportunity — which is exactly where licensing-plus-services deal structures become possible.
The licensing structure Japanese substrate firms will accept
Murata, Kyocera, and NGK do not view themselves as licensors. They view themselves as substrate manufacturers. Historically they refused to license LTCC IP to potential competitors even as the patents approached expiry, because the strategic position of being the only volume supplier was more valuable than royalty income.
That calculus is changing. Murata’s 2024 annual report noted structural pressure on substrate margins from Taiwanese and Korean competition. Kyocera divested its smartphone organic substrate line in 2023. The Japanese firms are now open to two structures they would have rejected five years ago.
The first is time-bounded technology transfer to non-competing geographies. A Malaysian OSAT producing LTCC substrate exclusively for ASEAN-assembled modules — not exporting bare substrates back into Japan, Taiwan, or Korea — is a sellable proposition. The Japanese substrate vendor preserves its core markets while extracting royalty income from a market it cannot serve cost-effectively anyway.
The second is joint-venture substrate fabs with retained Japanese process control. Kyocera has hinted at exactly this structure in conversations with Thai industrial estates. The Japanese partner owns the recipe, the local partner owns the building and the workforce, and the JV operates the line under shared governance. Singapore’s EDB has structured comparable JV vehicles for adjacent semiconductor technologies, and the template is portable to Penang or Eastern Seaboard Thailand.
The window and what closes it
Penang’s substrate gap is visible to everyone — including the Taiwanese substrate makers (Unimicron, Nan Ya PCB, Kinsus) who are also expanding into ASEAN. Taiwan moves faster on capacity decisions but holds weaker IP in the LTCC and HTCC categories specifically. The window for a Japanese-IP-based substrate fab in Malaysia or Thailand is open today and probably closes by 2028, when either Taiwanese capacity will have absorbed the volume demand, or the Japanese vendors will have set up their own ASEAN plants and removed the licensing option from the table.
The actionable position for an ASEAN industrial investor, a state-linked enterprise, or a Penang OSAT considering vertical integration is to open conversations now with Murata, Kyocera, and NGK Insulators about either licensing the expiring substrate IP or structuring a joint venture under one of the two acceptable templates. The Japanese substrate vendors are entering this decision window for the first time in their corporate history. Whoever moves first writes the terms — and substrate terms tend to lock in for a decade.
Interested in Japanese Deep-Tech for Your ASEAN Operations?
Technicity bridges Japanese bio-process innovation into ASEAN markets. Tell us about your needs and we will connect you with the right technology.