Malaysia's local semiconductor players are making their advanced packaging move, and the stakes could not be higher

A consortium of five Malaysian companies launched at SEMICON SEA 2026 with a target of 7% of the global advanced packaging market by 2035, an industry where the country currently holds no meaningful local position.

Every AI server, GPU and data center chip running today depends on a process called advanced packaging. It is how multiple chips–memory, logic, and compute–get integrated into a single high-performing unit. Without it, the AI infrastructure boom simply does not happen.

The global market for this technology was worth around US$51.6 billion in 2025 and is forecast to nearly double by 2031. For technology buyers and solution providers across Southeast Asia,

The question has shifted from what chips they are buying to where those chips are being assembled, and whether that supply chain is resilient enough to count on.

Five Malaysian companies—Inari Amertron, Pentamaster, NSW Automation, SkyeChip and startup FusionAP—formally launched the Malaysia Advanced Packaging Consortium, or MAPC, with an MOU signed in the presence of government leadership at SEMICON Southeast Asia 2026.

Their stated target: capture 7% of the global advanced packaging market by 2035, equivalent to an estimated US$5 billion in annual revenue. The challenge is the starting point. As Tan Eng Tong, a veteran of HP Labs and Seagate in Silicon Valley and former CEO of the Penang Skills Development Centre, put it plainly from the stage: "Malaysia is at zero. Nothing."

Why advanced packaging matters to the enterprise supply chain

The supply chain for advanced packaging is heavily concentrated. TSMC dominates the most sophisticated techniques from Taiwan. South Korea's memory giants, Samsung and SK Hynix, control high-bandwidth memory production. For enterprises and hyperscalers procuring AI infrastructure at scale, that concentration is increasingly a boardroom-level risk, not just a procurement footnote.

Southeast Asia has been identified as a critical diversification node. Malaysia's position in that picture is well established, but lopsided. The country accounts for roughly 13% of global chip assembly, test and packaging output. The problem, as MAPC's own presentation made clear, is that this 13% sits almost entirely within multinational facilities. Intel, Micron, Infineon and others operate the advanced work here. Local Malaysian companies own none of that process knowledge.

Andrew Goh, Corporate Vice President and General Manager for Southeast Asia at Lam Research, one of the world's leading semiconductor process equipment suppliers, offered a slightly different read in a separate interview at SEMICON SEA–though the conclusion is the same.

"If you look at today, Malaysia on advanced packaging, from zero to 100%, we are probably about 10%," he said. That 10% largely reflects what multinationals are already doing on Malaysian soil. What Tan was pointing to is the share owned by local companies, and on that measure, the number is closer to what he said: nothing.

Goh’s view was that advanced packaging represents the most realistic near-term path for Malaysia to move up the semiconductor value chain, precisely because competing with Taiwan on front-end wafer fabrication is no longer a credible ambition. "Find an area where you can be complementary," he said. "Advanced packaging, I think it's not too late to enter."

What the consortium is actually building

MAPC's initial project is an HBM4 pilot line—HBM4 being the next generation of high-bandwidth memory stacking that AI accelerators require. Each of the five founding members contributes a defined capability: SkyeChip on IC design, Inari on OSAT production infrastructure, FusionAP on packaging process development, and Pentamaster and NSW Automation on testing and equipment respectively.

The structure took two years to define before the MOU was signed. IP-sharing rules, milestone schedules and project responsibilities are already agreed. Noorazidi Che Azib, Vice President of Inari Amertron Bhd, described the approach as a shift rather than a leap—using autonomous manufacturing and advanced packaging to move up within the manufacturing segment, rather than attempting a wholesale reinvention into IC design or wafer fabrication overnight.

FusionAP, the startup at the centre of the pilot line ambition, was founded by former Intel advanced packaging executives who chose Malaysia specifically because the existing talent base and ecosystem made it viable. Co-founder Peter Charvat, who spent years at Intel working on advanced packaging, described the consortium model as the only credible path for companies at this stage: "Advanced packaging cannot be done by a single company alone. Moving too fast, too complicated, no standards."

The questions the market will be watching

The MAPC launch came with a frank acknowledgement of what remains unresolved. When asked about funding for FusionAP's pilot line, independently estimated by Hong Leong Investment Bank at RM400 million, with a further RM2 billion required to scale to mass production, Charvat confirmed that the capital has not yet been secured. "No, we do not have RM400 million. We're going to have to work it out." The plan is to acquire equipment incrementally as customer engagements develop, on a lease-to-own basis where possible.

The anchor customer question is equally open. As one attendee from Vitrox put it directly to the panel: without a committed customer to validate the supply chain, how does confidence build? Tan's response was pragmatic: you cannot attract a customer without demonstrating capability, and you cannot build capability while waiting for a customer.

The consortium model exists precisely to break that deadlock by letting five companies share the cost and risk of proving the technology before commercial scale is required. HLIB's analysis of MAPC and FusionAP, published in March 2026, noted that advanced packaging commands gross margins of 40 to 50%, compared with 15 to 20% for traditional packaging.

The financial case for success is compelling. Execution, in a segment where the global technology frontier moves every twelve months, is the variable. Lam Research's Lee Chee Ping, Managing Director for Strategic Marketing in Advanced Packaging, framed the time pressure clearly: "Nvidia's product cadence has shortened from two to three years per product to every year, or more. That pressure is high for every player in the semiconductor ecosystem to design and innovate faster."

What this means for the regional technology ecosystem

For solution providers and technology buyers across Southeast Asia, MAPC's progress, or lack of it, will shape where AI hardware supply chain risk sits over the next decade. A credible local advanced packaging capability in Malaysia would reduce the region's dependence on Taiwan-concentrated supply for the chiplet-based AI accelerators that are increasingly central to enterprise and cloud infrastructure procurement decisions.

The window is real. Intel's advanced packaging complex in Penang, 99% complete and backed by a RM12 billion capital commitment, is slated to begin assembly and testing operations later this year. That brings front-end-adjacent advanced packaging process knowledge onto Malaysian soil for the first time at commercial scale, knowledge that a consortium like MAPC could, over time, help diffuse into locally owned companies.

Whether the National Semiconductor Strategy, which has advanced packaging as a stated priority, translates that ambition into the specific investment and infrastructure MAPC needs to execute is the question the next twelve months will begin to answer.

The target is 7% of a US$90 billion market by 2035. The starting point is zero. The distance between them is large, but the direction, for the first time, is clearly set.