This article first appeared in Digital Edge, The Edge Malaysia Weekly on September 14, 2026 - September 20, 2026

Malaysia’s bid to capture 7% of the global advanced packaging market by 2035 is beginning with five local companies developing an advanced packaging prototype built around an HBM4-capable chip for artificial intelligence (AI) and high-performance computing applications — in a collective attempt to help locally owned semiconductor players get a foot in the door.

The five companies are chip designer SkyeChip Bhd, advanced packaging start-up FusionAP Sdn Bhd, outsourced semiconductor assembly and test company Inari Technology Sdn Bhd, and equipment makers NSW Automation Sdn Bhd and Pentamaster Instrumentation Sdn Bhd. Together, they form the Malaysia Advanced Packaging Consortium (MAPC).

The 24-month project, which began in February, is unprecedented in Malaysia in terms of funding scale and industry collaboration. It has received RM92 million in R&D grants from the government and RM93 million from the participating companies.

The government funding is provided through the Malaysia Science Endowment, which comes under the Ministry of Science, Technology and Innovation (Mosti), with the Academy of Sciences Malaysia (ASM) as the endowment’s implementing agency.

Malaysia already accounts for about 13% of global semiconductor assembly, testing and packaging, and has more than 50 years of manufacturing experience, but the capabilities of local companies remain fragmented across parts of the value chain, leaving the country without an end-to-end advanced packaging chain.

At a time when advanced packaging has become increasingly important to the next generation of computing, and as it becomes harder and more expensive to shrink transistors, chipmakers are combining processors, memory and other specialised dies in extremely sophisticated packages that will allow these to operate as a single system.

Malaysia risks seeing its share of the semiconductor packaging market shrink if it continues to rely on conventional packaging and fails to move into newer areas such as advanced packaging, stress industry pundits.

The global advanced chip packaging market is expected to double from US$43.8 billion in 2025 to US$87.6 billion by 2030, representing an annual growth of 14.8%, according to US-based market research firm BCC Research.

The 2.5D and 3D packaging segment could grow even faster, at about 21% annually, driven principally by AI, high-performance computing and HBM integration, said Kenanga Research in a June report.

But for Malaysia to capture a meaningful share of the market, it will need to develop and own the process knowledge and intellectual property behind advanced packaging, experts tell Digital Edge.

That will require domestic sources of specialised equipment and materials, a deeper pool of engineers and stronger links between researchers and industry, while local providers will have to show that their processes can meet the cost, quality and reliability standards demanded by customers. Advanced packaging also requires the chip, package, equipment and manufacturing processes to be developed together.

Boo Kiang Ngee, founder and managing director of NSW Automation, says the complexity of advanced packaging means no single company can cover the entire process on its own.

“There is a lot of equipment involved. There are a lot of processes involved. So one company cannot do the job … Local companies can collaborate with each other and complement each other to make our ecosystem more robust,” he explains.

Industry veteran Tan Eng Tong, convenor of the MAPC, says the country already has companies operating across different parts of the semiconductor value chain, but those capabilities have largely been developed separately. He was previously vice-president of strategic management at SilTerra Malaysia Sdn Bhd, where he worked on the foundry’s expansion into areas such as silicon photonics, life sciences and automotive applications.

The MAPC is an attempt to bring those capabilities together in an integrated production chain. But getting companies to work together is one of the project’s biggest challenges, says Tan.

“The government cannot force companies to work together. The companies have to decide that they want to work together and they have to trust each other. That is actually the most difficult part,” he adds.

Tan points to Taiwan’s Industrial Technology Research Institute (ITRI) as an example of the kind of collaborative model Malaysia needs to develop. Taiwanese semiconductor companies have long been accustomed to working together to develop technology and build capabilities, before taking what they have learnt back to their own businesses and competing with one another, he says.

ITRI, which was established in 1973, has played a central role in building Taiwan’s semiconductor industry by developing technologies, transferring them to industry and creating companies around them. It spun off United Microelectronics Corp in 1980 and later transferred facilities, technology and engineers to help establish Taiwan Semiconductor Manufacturing Co (TSMC) in 1987.

ITRI has continued to provide shared research and pilot platforms where semiconductor companies can collaborate on technologies such as heterogeneous integration and advanced packaging before commercialising them individually.

The MAPC is an attempt to replicate that collaborative model among the country’s foremost semiconductor firms, based on the premise that competition has evolved from company versus company to supply chain versus supply chain. The five companies have been given a common project that requires them to combine their respective strengths, understand what each can contribute and develop capabilities they can later take to other customers.

The first test is an advanced-packaging prototype built around an HBM4-capable chip, with each company responsible for a different part of the process and the output of one feeding into the work of the next. Working on the prototype will allow each company to understand what the others can do and where they could continue collaborating after the project is completed, says Tan.

The prototype is not intended to be a commercial product. It is a proof of concept or a way to demonstrate that the different technologies, equipment and processes developed by the five companies can be connected successfully to produce a working advanced package, he says.

Malaysia is already playing catch-up in advanced packaging, where technologies are advancing quickly. For example, TSMC’s CoWoS 2.5D packaging platform entered commercial production in 2012 and has continued to evolve since, while Malaysia is only now attempting to establish an integrated advanced packaging capability among locally owned companies.

The prototype, if successfully built and validated, would give the five companies proven processes, equipment and know-how that they can take to prospective customers. The bigger test is whether the project can extend those capabilities beyond the original five collaborators.

In a written parliamentary reply in June, Mosti said the five companies were working with local universities and research institutions on integrated R&D. The ministry also identified advanced packaging as a national priority under its Mission-Oriented Initiative for Advanced Semiconductors.

Digital Edge asked ASM and Mosti about the wider MAPC programme, including how the project will be implemented, how its benefits will extend beyond the five companies and what will come after the initial prototype.

ASM, which was coordinating the response with Mosti, says it is unable to provide further input at this stage as several rollout actions and regulatory processes are still being addressed.

Tan assures that the MAPC project is intended to continue beyond the first prototype, using each project as a starting point rather than an end in itself. The HBM4-class prototype project will give the five companies experience working together as a chain, and once those capabilities and relationships have been built, they could be used for commercial work or other technologies. The companies have already begun discussing how they could continue working together after the prototype is completed, he adds.

The consortium is also intended to grow beyond the original five companies, with future projects potentially bringing together different groups, depending on the technology and capabilities required. It is already looking at silicon photonics as the next area it wants to pursue, followed by co-packaged optics, with both developed progressively depending on the companies and capabilities involved.

Other areas, such as power semiconductors, silicon carbide devices and sensors, could eventually form the basis of separate advanced packaging projects. The wider effect could also extend through the companies’ existing supply chains.

Each of the five companies brings its own network of suppliers and partners. At a workshop last year, each of them brought three or four companies from their ecosystem, resulting in about 15 to 20 companies participating.

As the companies move into more sophisticated packaging, their equipment, materials and process suppliers may also have to develop new capabilities. For example, NSW’s Boo says the company is looking to work with local producers because many of the materials required for advanced packaging are still sourced from overseas. These include substrates, thermal-interface materials, adhesives and underfill materials.

A local supplier could bring a material into the MAPC process to be tested and validated, giving it a chance to refine the product until it meets the standards required by customers. This could give Malaysian materials companies an entry point into a market currently dominated by overseas suppliers, he says.

“If a local company wants to start this, of course, it will require a certain level of investment as well as time and patience. But the important thing is there is an opportunity for the local company to start somewhere,” he adds.

Ooi Teng Chow, co-founder and CEO of FusionAP, says the company has spoken to other start-ups about using its test chips for their own testing.

Asrulnizam Abdul Manaf, associate professor and director at Collaborative Microelectronic Design Excellence Centre at Universiti Sains Malaysia, reveals that FusionAP approached his team to discuss co-developing a process design kit (PDK), while Inari separately asked whether USM could help develop its assembly design kit (ADK). He says the ADK would capture the design rules and process capabilities of Inari’s packaging line, such as the minimum bump pitch and redistribution layer requirements, so chip designers such as SkyeChip can design packages around what the line can manufacture.

Moreover, universities can support the companies with fundamental research, modelling and process development, including work on materials, package and interposer design and the design kits needed to turn manufacturing processes into repeatable production, says Asrulnizam.

Universities also play a role in developing the engineers needed to sustain the industry. USM has introduced a master’s degree programme in advanced semiconductors and packaging covering areas such as redistribution layers, interposers, modelling and simulation.

Nevertheless, Asrulnizam says universities cannot realistically afford to build and maintain their own complete 2.5D or 3D packaging lines.

That is why he is proposing that the MAPC companies open their production lines to postgraduate students and researchers through internships, joint research projects and postgraduate programmes. This would give students hands-on experience in areas such as interposer development and bonding processes, while allowing the companies to tap university research and build a pipeline of engineers with practical advanced packaging experience.

The effort to bring Malaysian companies together around advanced packaging R&D began about two years ago.

Mimos, Collaborative Research in Engineering, Science and Technology (CREST) and ASM brought government agencies, universities and semiconductor companies into the discussions, with one of the early meetings held at Universiti Sains Malaysia, says Tan. However, participation began to thin when the discussions moved from drawing up the broader road map to the question of “who’s going to do what”, and the effort eventually stalled.

Tan subsequently worked with CREST to restart the discussions. When the industry players were brought together again, five companies came forward, each with its own proposal for what it could develop.

The turning point came when ASM president Academician Datuk Dr Tengku Mohd Azzman Shariffadeen pushed the companies to turn the five proposals into one project, with a common outcome rather than five technologies being developed separately.

Tan says Tengku Azzman told the companies that the funding would only follow if they turned the proposals into one mission — a project with a clear beginning, middle and end, and where the output of one company connected with the work of the next. The companies then reworked their individual plans around the common advanced packaging project and calculated what each would need to spend to deliver its part.

The RM185 million funding is not divided equally among the five companies. Each company’s allocation is based on the work it has committed to perform and the amount it expects to spend, with the government providing matching grants.

The matching structure also gives the companies a strong financial incentive to deliver because each participant is putting its own resources into the MAPC alongside the government funding, says Tan.

He adds that the individual commitments run into the tens of millions of ringgit, giving the companies significant “skin in the game”. “When a company has committed RM10 million, RM30 million, this is a serious endeavour. They intend to make it successful.”

The government funding is released upon agreed milestones and deliverables and each company’s expenditure is subject to audit.

Only SkyeChip’s allocation is publicly known. The company was the latest among the five to list on Bursa Malaysia on May 20. Its prospectus states that ASM, acting on behalf of Mosti, awarded the company RM12 million on March 4, 2026, principally for the design and development of test chips through tape-out.

The advanced packaging prototype package starts with SkyeChip, which is providing the test vehicle around which the rest of the line is being developed. The company is designing the test chip, silicon interposer, package and board, while procuring the HBM stack memory that will sit alongside its chip.

The test chip and interposer will be fabricated by third parties before the components are passed on for assembly. Once the completed package comes back, SkyeChip will carry out electrical characterisation to determine whether it functions and how its performance compares with an existing proven advanced packaging process.

Its CEO Datuk Fong Swee Kiang says the chip, interposer and package can no longer be treated as separate designs because electrical signals, power, heat and mechanical stresses interact across several dies within the same package.

Heat can cause the thin interposer to bend and affect the tiny connections between the chips. Therefore, SkyeChip has to design the package as a whole, taking into account its electrical, thermal and physical behaviour, while relying on third parties for fabrication and components.

“So we are acting like enabler, but at the same time also as control. So when you need to design an experiment, you need a control and you need the experiment to see what is the difference. So when we put our HPM4 test chip, we actually design the interposer and design the package and to be assembled in a proven OSAT,” says Fong.

From there, the components move to FusionAP, which is responsible for the 2.5D integration. The company will take SkyeChip’s chiplet and the procured HBM memory and stack them onto the silicon interposer — a process known as die-to-wafer integration.

For the first prototype, this stage will be carried out at the facility of FusionAP’s technology partner in Europe before the integrated stack is brought back to Penang. Its technology partner is Belgian semiconductor research centre imec (Interuniversity Microelectronics Centre), one of the world’s leading semiconductor R&D institutes. Based in Leuven, imec works with chipmakers, equipment suppliers and other companies to develop and test new semiconductor technologies using its research facilities and pilot lines.

FusionAP brings its advanced-packaging know-how, while its technology partner provides the IP and existing facility needed for the initial integration, says Ooi.

The company’s MAPC allocation is being used for R&D, technology development and technology transfer instead of buying equipment or building a production facility. This includes licensing intellectual property and paying to use the equipment at its technology partner’s facility, while FusionAP develops and validates the process, says Tan.

One of the immediate difficulties is integrating HBM4 with the other components on the interposer, says Ooi.

Differences in the characteristics and thickness of the materials can cause the structure to warp or crack, requiring close control of the materials, temperature and process formula. Therefore, the companies are starting with cheaper mechanical samples before putting the actual SkyeChip components through the process.

The longer-term challenge for FusionAP is to bring that capability to Malaysia. The process being used at its technology partner’s facility is closer to a laboratory-scale operation, and the company will eventually have to transfer the know-how, automate the process and make it repeatable enough for commercial production, says Ooi.

Doing so will require significantly more capital.

Hong Leong Investment Bank Research estimated that FusionAP would need about RM400 million for a pilot production line and a further RM2 billion to scale up to high- volume manufacturing, reported The Edge Malaysia on March 25. These investments would be separate from the MAPC investments.

Once the dies have been integrated on the interposer, the stack returns to Penang and moves to Inari, where it will be attached to the substrate and completed into the final package. The company will put together the pilot production line at its Batu Kawan facility, building on parts of its existing line and adding equipment developed by NSW Automation and Pentamaster.

NSW Automation is developing equipment for underfill, thermal interface material and adhesive dispensing, as well as handling and attaching the package lid. While the company already works with some of these processes, Boo says the MAPC machines are new models designed around the package’s different form factor, including a new equipment platform and handler.

Pentamaster will provide the testing and inspection equipment for the Inari line, including X-ray and metrology systems used to check the package during assembly and determine whether it meets the required specifications. The immediate target is to have the equipment for the pilot line installed at Inari by end-2026, says Tan.

Once installed, the companies will have to test and fine-tune the equipment and processes until the production flow is stable and repeatable.

In 2027, the actual SkyeChip test chip, interposer and high bandwidth memory (HBM) will have to begin moving through the process.

FusionAP will first integrate the chiplet and HBM onto the interposer in Europe before the completed stack is sent back to Penang for final packaging at Inari. FusionAP is targeting the completion of the actual SkyeChip package by the third quarter of 2027.

The finished package will then be sent back to SkyeChip for electrical testing to determine whether it works and how its performance compares with the benchmark. Any problems identified will be fed back to the other companies so they can refine their equipment and processes.

The remaining period will be used for testing, validation and further refinement, with the 24-month MAPC programme scheduled to end in February 2028.

Save by subscribing to us for your print and/or digital copy.

P/S: The Edge is also available on Apple's App Store and Android's Google Play.