- Five local companies form consortium to build what Malaysia has never had.
- No funding secured, no anchor customer, and the clock is running.
- SEMICON SEA 2026: local players still don’t own an advanced packaging process.
The number that stopped the room at MITEC on Wednesday afternoon was not the target. It was the starting point.
“Malaysia is at zero,” said Tan Eng Tong, a veteran of HP Labs and Seagate in Silicon Valley, former CEO of the Penang Skills Development Centre, and one of the sharper industry voices in the room – representing the newly formed Malaysia Advanced Packaging Consortium, or MAPC, from the main stage of SEMICON Southeast Asia 2026.
The target Tan was referring to is audacious: 7% of the global advanced packaging market by 2035, worth an estimated US$5 billion annually if achieved. The consortium behind it, five founding members that include Inari Amertron, Pentamaster, NSW Automation, SkyeChip and startup FusionAP, formalised their partnership that same afternoon in front of the country’s Deputy Minister of Investment, Trade and Industry (MITI), Sim Tze Tzin, and a packed audience of industry leaders, policymakers and investors.
The MOU was signed. The photo was taken. But before the applause had faded, the harder questions were already in the room.
What advanced packaging is and why Malaysia doesn’t have it
To understand what MAPC is trying to build, it helps to understand what it’s trying to close. Conventional semiconductor packaging, the kind Malaysia has used for five decades, involves encapsulating a single chip in a protective casing, with wire bonds connecting it to a circuit board. It is essential work. Malaysia’s OSATs (outsourced semiconductor assembly and test houses) account for roughly 13% of global output in this category.
That is no small thing – but advanced packaging is a different discipline entirely. Instead of housing one chip, it integrates multiple chips – sometimes from different manufacturers, different processes, different functions – into a single, unified package. Memory sits beside the processor. Logic tiles stack on top of each other. Optical components share a package with conventional silicon.
The result is dramatically higher performance and energy efficiency than what a single chip, no matter how advanced, could deliver alone. It is what makes today’s AI accelerators work, and demand for it is growing faster than the industry can currently supply. As Tan put it bluntly, using his own slide: “Malaysia OSATs ship 13% of global chips. Mostly by MNC facilities located here. The advanced techniques are really in the MNCs. The local companies really don’t have any.”
That distinction matters enormously. Malaysia’s 13% is real and valuable, but it is held by Intel, Infineon, Micron, and others who have planted their fabs and assembly lines here. Malaysian companies with Malaysian process knowledge do not own it. If those MNCs decide to change operations, say to India, to Vietnam, or to wherever the next incentive package is more compelling, that 13% goes with them.
The consortium model: why five companies decided they couldn’t wait
The logic behind MAPC, as Noorazidi Che Azib, Vice President of Inari Amertron Bhd, explained – and Tan echoed – is that no single local company has the capital, expertise or production volume to develop advanced packaging ability alone. Advanced packaging moves too fast, costs too much, and has no settled standards. As MAPC’s own slide stated:”By the time you buy the latest equipment – if you are allowed to – you are out of date.”
The consortium’s initial project is to develop an HBM4 pilot line: high-bandwidth memory, being the type of specialised memory stack that AI chips require to function at scale. Each of the five companies plays a defined role: SkyeChip handles IC design, Inari contributes its existing OSAT production lines, FusionAP leads on packaging process development, while Pentamaster and NSW Automation cover testing and automated equipment, respectively.
The work took two years of internal discussion to define before the MOU was signed. The project structure, milestone schedule and IP-sharing rules are already agreed. Inari’s Noorazidi was characteristically direct about why Inari chose to join not go it alone: “Moving up the value chain is not that easy. It is something that we want to do, but it’s not easy with respect to cost and ability. However, it’s not impossible.”
His Inari slide, showing the smiley curve of semiconductor value creation, made the point visually – Malaysia’s local players sit at the bottom of the curve, in manufacturing, while R&D and product design at either end command the margins.
The ambition, as Tan summarised it, is for “one plus one equals five”. Alone, each company hits a ceiling. Together, with shared process development, shared talent pipelines and protected IP arrangements, the ceiling moves.
The questions not answered
The MOU ceremony and the consortium presentations were the easy part of the afternoon. The harder part came during the open Q&A.
When asked directly about the funding status 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 – Peter Charvat, co-founder of FusionAP, confirmed what the room suspected. “No, we do not have RM400 million. That’s a good question. We’re going to have to work it out.”
His plan, as he outlined it, is to acquire equipment incrementally as customers come in, ideally on a lease-to-own basis. It is a reasonable strategy for a startup. It is also a constraint for a project that requires front-end-grade process equipment – machines that cost millions of dollars each, with lead times now running 12 months or longer industry-wide.
The funding question was followed by another that cut closer to the structural problem: where exactly on the advanced packaging technology curve is Malaysia actually entering? Charvat was candid: FusionAP is targeting silicon interposer-based 2.5D packaging for its initial phase, with co-packaged optics in the next four to five years.
The consortium is not starting with hybrid bonding, the most sophisticated form of chip stacking currently in commercial production, because the ability and capital requirements to go there immediately are out of reach.
Tan’s response to this acknowledged the gap plainly. “I said Malaysia is starting from zero. So yes, the answer is zero.” His argument, made forcefully, was that zero is the honest starting point from which credible progress can be planned, not a reason for paralysis. “If we don’t start, it’s not going to happen.”
The 50-year question
The session’s most substantive exchange came when the panel was pressed on why, after five decades of semiconductor presence in Malaysia, local players have not developed this ability already. The answers were honest in a way that policy forums rarely allow.
Amarjit Sandhu, Corporate Vice President for Assembly, Test and NAND Operations at Micron Memory Malaysia, one of the largest MNC operators in Malaysia with over 5,000 employees in Penang and close to RM2 billion in annual local supplier spending, put it plainly: the entire global supply chain was caught unprepared by the speed of AI’s emergence.
“No one anticipated the AI transformation to happen as fast as it happened. Everyone in the supply chain was literally caught with their pants down. No capacity, processes not mature, equipment sets not ready,” Sandhu noted. That is a global condition, not a uniquely Malaysian failure.
But there is a structural piece that runs deeper. The FDI model that brought semiconductor investment to Malaysia has always been optimised for volume, cost efficiency and geopolitical stability, not for local knowledge transfer into process engineering. MNCs invest in what they need, not in what Malaysia needs to develop next. Heng-Jan Jonge Poerink, Senior Vice President for Global Operations and Managing Director of BESI APAC, has operated in Malaysia for sixteen years and builds the majority of the company’s advanced hybrid bonding equipment in Shah Alam with 80% local supply chain content.
His read on the challenge was different: “Companies like ASML, Lam, Applied, KLA – they all look for suppliers who can supply at the module level. For module level, you need to have your purchasing, engineering, machining, cable harnesses, and clean room under one roof.” Most Malaysian SME suppliers are not there yet. Some are close. The gap is real, but it is also close-able.
What is running out, faster than capital or technology access, is time. As Micron’s Sandhu noted, Nvidia is now releasing a new AI chip product every twelve months. That cadence compresses every supplier’s reaction window. “As we run faster, we want our suppliers to run faster as well. Run faster, with no mistakes.”
India is scaling its semiconductor ambitions quickly. Vietnam is already absorbing manufacturing work that was once considered Malaysia’s natural territory. The window that geopolitical supply chain diversification has opened for Malaysia – being a neutral, experienced, English-proficient manufacturing hub at a moment when the world wants alternatives to Taiwan and China concentration – will not stay open indefinitely.
What MAPC needs to succeed
Tan, Inari’s Noorazidi and Micron’s Sandhu all converged on the same point when pressed: the change in mindset required from local industry is as important as any policy instrument or investment incentive. “You have to run faster by first knowing what Nvidia is doing – you can read it. If you can’t, go to AI. All the information is available,” said Tan. “Don’t wait for Micron to tell you that you need micro bumps. You should know how to do that.”
That is the model MAPC is trying to prove. Not waiting to be told what the market needs, but building toward it proactively, from a consortium that took two years to define its own project before announcing it publicly. The five founding companies are not short on credibility: Inari is Malaysia’s largest listed technology company, Pentamaster is a respected instrumentation player, and SkyeChip brings IC design ability that few local companies have.
FusionAP’s founding team includes veterans of Intel’s advanced packaging operations who chose to come back to Malaysia specifically because the ecosystem and talent base here made it viable. The honest challenge, as HLIB’s March 2026 analysis noted, is that MAPC and FusionAP are still in the very early stages of execution. Securing funding, landing an anchor customer, and proving out the pilot line are all milestones that remain ahead.
Advanced packaging commands gross margins of 40 to 50%, compared to 15 to 20% for traditional packaging. The prize is real. So is the distance to it. The government was in the room. The MOU was witnessed by the minister. The National Semiconductor Strategy has advanced packaging as a stated priority.
Whether that translates into the specific, committed and sustained support – R&D infrastructure, equipment access, talent development at the process engineering level – that a consortium like MAPC actually needs to execute is the question that the next twelve months will begin to answer. Zero is not a position anyone can afford to hold for much longer.
Dashveenjit Kaur covered the MIDA Strategic Panel Forum at SEMICON Southeast Asia 2026 at MITEC, Kuala Lumpur, on 6 May 2026. Read her interview with Lam Research’s Andrew Goh and Lee Chee Ping on Malaysia’s advanced packaging semiconductor opportunity.
Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information.
Tech Wire Asia is powered by TechForge Media . Explore other upcoming enterprise technology events and webinars here .