(Sept 17): When Intel opened its first overseas manufacturing facility in Penang in 1972, Malaysia’s role was largely to assemble chips designed and fabricated elsewhere. More than half a century later, the company is completing an advanced packaging complex in the same state. Advanced packaging uses sophisticated manufacturing processes to integrate multiple specialised chips into a single package.
That shift captures both Malaysia’s gains from the artificial intelligence (AI) boom and its unfinished industrial challenge. Advanced packaging can bring more sophisticated technology and skilled jobs. But if the products, process technologies, intellectual property, and customer relationships remain predominantly foreign-owned, Malaysia may capture less long-term value than the investment headlines suggest.
Asia’s AI hardware supply chain is highly concentrated. US companies design many leading accelerators — the specialised processors that perform the heavy calculations required by AI — while Taiwan fabricates and packages many of them, and South Korean supplies much of the high-bandwidth memory paired with them.
Yet the accelerator cannot operate alone. It requires memory, storage, high-speed networking, optical connections, power-management chips, cooling systems, servers, and data centres. Malaysia participates in many parts of this wider ecosystem, while investment by the world’s largest cloud-computing companies is driving data center construction, particularly in Johor. Its AI gains have therefore come mainly from the hardware surrounding the accelerator, rather than from making the accelerator itself.
Advanced packaging offers one of Malaysia’s clearest opportunities to move closer to the performance-defining core of this supply chain. As transistor scaling becomes harder and more expensive, chipmakers increasingly improve performance by combining processors, memory, and specialised chips within a single package. Shorter connections allow data to move faster while consuming less power — an essential requirement for AI.
This is far more complex than placing a finished chip inside a protective casing. Advanced packaging requires microscopic wiring, precision bonding, thermal management, and exact alignment across multiple dies. It resembles operating a miniature chip factory more than a conventional assembly line.
Intel’s Penang project illustrates the scale of this technological shift. Announced in 2021 as part of a US$7 billion (RM28.7 billion) investment in Malaysia, the complex contains a 710,000 sq ft cleanroom. Intel’s advanced packaging portfolio includes EMIB, which connects chiplets within a package, and Foveros, which stacks them vertically. After delays, the complex’s first phase is expected to begin advanced packaging operations later in 2026.
Intel is not alone. Taiwanese firms are also moving in the same direction, although their announced Malaysian capabilities are at a different technology tier. Chipbond opened a Penang facility in February 2026 offering advanced wafer bumping and wafer-level chip-scale packaging — processes that add microscopic electrical contacts and package chips before the wafer is cut into individual dies. SPIL is developing a major packaging-and-testing facility expected to introduce similar capabilities, while ASE has continued expanding its long-standing presence in Penang.
Together, these investments can move Malaysia’s production frontier. More sophisticated manufacturing, process engineering, and testing will be performed locally, creating skilled employment and deeper industrial capabilities.
But upgrading will remain predominantly foreign-led. Multinational companies generally own the products, process technologies, intellectual property, and major customer relationships. Malaysia can therefore move up the semiconductor value chain without necessarily retaining a proportionate share of the resulting profits or strategic control.
The economic question is not merely where chips are packaged. It is whether foreign investment raises domestic productivity, wages, technological capability, and domestic value added. That depends on how deeply multinationals embed higher-value activities in Malaysia and how far capability diffuses beyond their fence line — pointing to two complementary policy tracks.
The first is to deepen the role of multinational companies. Investment incentives should be tied not only to capital expenditure but also to process engineering, package design, product development, workforce training, research collaboration, and the qualification of Malaysian suppliers. The objective is to make Malaysian operations technologically important within each company’s global network, and consequently harder to relocate.
The second track is to build capabilities that Malaysian firms can own. The most credible opportunities lie in specialised chip and package design, testing equipment, automation, inspection systems, precision components, factory software, and selected materials. These activities may attract fewer headlines than a leading-edge fabrication plant, but they offer greater potential for domestic ownership of technology and intellectual property.
The Malaysia Advanced Packaging Consortium is a promising experiment in this direction. Its members — SkyeChip, Inari Technology, FusionAP, Pentamaster Instrumentation, and NSW Automation — combine capabilities in chip and chiplet design, semiconductor assembly, automated testing, precision equipment, and advanced packaging process development.
For now, however, the consortium remains an early-stage platform, not evidence that Malaysia already possesses a home-grown frontier capability. Success will depend on whether collaboration produces customer-qualified processes, defensible intellectual property, and paying customers.
The link between the two policy tracks is crucial. Spillovers do not occur merely because foreign and domestic companies occupy the same industrial park. They arise when Malaysian engineers acquire advanced production experience, local companies co-develop solutions with multinational customers, and domestic suppliers qualify for global programmes.
Malaysia’s success should not be judged by whether it eventually manufactures a leading AI accelerator. The more immediate test is whether it can leverage foreign investment to bring advanced packaging, product engineering, and research into the country — and connect those activities to Malaysian-owned design, equipment, automation, and testing companies.
Intel can deepen the technological sophistication of production in Malaysia; the consortium is an attempt to increase the share of that technology Malaysian firms eventually own. Together, they embody the two tests of successful upgrading: how advanced production in Malaysia becomes, and how much value Malaysia retains after the AI investment cycle turns.
Wee Chian Koh is a senior economist at AMRO and Kian Heng Peh is its group head and lead economist.