Malaysia's Semiconductor Push: Reading the Signals

Global chipmakers keep announcing Malaysia expansions. The useful question is which announcements build durable capability, and which the country's own talent numbers say it cannot yet fully staff.

A record year, by the numbers MIDA has published

Malaysia’s electrical and electronics sector secured RM28.5 billion in approved investments in 2025, a figure the Malaysian Investment Development Authority (MIDA) has cited as evidence the country’s semiconductor position is entering “its next phase of growth from a position of strength” (MIDA, SEMICON Southeast Asia 2026 media release). That headline sat behind the return of SEMICON Southeast Asia to Kuala Lumpur’s MITEC in May 2026, and it was followed within weeks by a run of individual project announcements that gave the aggregate figure some texture.

The clearest of these was AIXTRON SE’s agreement with MIDA, signed at the SEMICON SEA event on 5 May 2026, to build a greenfield deposition-equipment facility on 8.5 acres at Bandar Cassia Technology Park in Penang (MIDA media release, 5 May 2026; The Star, 5 May 2026). AIXTRON’s CEO Felix Grawert described the decision as reflecting “long-term confidence in Malaysia as a strategic semiconductor location,” and MIDA’s own framing was more pointed: chief executive Datuk Sikh Shamsul Ibrahim Sikh Abdul Majid said the deal signalled Malaysia “moving beyond assembly-centric activities to become a global hub for the critical enablers of the artificial intelligence and electric vehicle revolutions.”

The distinction the aggregate figure hides

That framing is worth testing against what actually arrived. Two other announcements from the same period illustrate a real distinction. Chipbond Technology Corporation opened a roughly US$200 million (about RM800 million) advanced packaging and testing facility in Batu Kawan, Penang, in February 2026, offering wafer bumping, wafer-level chip-scale packaging and flip-chip assembly and test (MIDA media release, 9 February 2026). In June 2026, MKS Inc opened a Penang factory representing over RM400 million in investment, officiated by Prime Minister Anwar Ibrahim (The Edge Malaysia, 20 June 2026). Both are genuine, capital-intensive investments. Both are also, by MIDA’s own description, outsourced semiconductor assembly and test (OSAT) work — the packaging and testing layer of the value chain, not front-end wafer fabrication or chip design.

AIXTRON is a different kind of signal. It is not a chip manufacturer opening a back-end plant; it is a maker of the deposition equipment that front-end fabrication depends on, choosing to locate manufacturing capacity in Malaysia rather than simply selling equipment into the country from elsewhere. That is a step up the value chain, and it is why MIDA’s own commentary leaned on it so heavily. The distinction matters for anyone using investment headlines to judge how durable Malaysia’s semiconductor position actually is: OSAT capacity is real employment and real capital, but it is also the segment most exposed to relocation if costs shift, because the barriers to moving packaging and testing elsewhere are lower than the barriers to moving equipment manufacturing or fabrication.

Where Malaysia is still starting from zero

The clearest evidence that Malaysia has not yet closed this gap came from its own industry voices. At the same SEMICON SEA 2026 event, Tan Eng Tong — a veteran of HP Labs and Seagate, representing the newly formed Malaysia Advanced Packaging Consortium (MAPC) — told the audience plainly: “Malaysia is at zero” in advanced packaging capability (TechWireAsia, 8 May 2026). MAPC’s five founding members — Inari Amertron, Pentamaster, NSW Automation, SkyeChip and FusionAP — set a target of capturing 7% of the global advanced packaging market by 2035, worth an estimated US$5 billion annually if achieved. Advanced packaging carries gross margins of 40–50%, against 15–20% for traditional packaging, which is precisely why the ambition exists — and precisely why starting from zero is a meaningful admission from the people closest to the work.

The talent constraint is not a footnote

The more binding constraint may not be capital at all. According to reporting from the same event, the Malaysian government has acknowledged the industry needs 50,000 skilled engineers to meet current demand, while Malaysian universities produce approximately 5,000 engineering graduates annually — a tenfold gap that, as the reporting notes, “is not one that any policy can bridge quickly” (TechWireAsia, 5 May 2026). The National Semiconductor Strategy targets training 60,000 highly skilled engineers by 2030, and initiatives such as the Penang STEM Talent Blueprint and CREST are underway. But the same reporting cites Wong Siew Hai, president of the Malaysia Semiconductor Industry Association, telling the South China Morning Post roughly a year earlier that the country loses an average of 15% of its semiconductor talent annually to brain drain.

Compounding this, revised Employment Pass salary thresholds took effect on 1 June 2026: Category III, which covers technical specialists and semiconductor manufacturing technicians, moved from a range of RM3,000–4,999 to RM5,000–9,999, with a manufacturing-sector floor of RM7,000, alongside new requirements for formal succession plans for foreign hires in this category. The intent — pushing employers toward training and retaining local talent rather than relying indefinitely on foreign technical staff — is defensible policy. The near-term effect, for companies already short-staffed, is a real cost increase precisely as investment commitments are accelerating.

What this means for a market-entry or supply-chain decision

None of this argues against Malaysia’s semiconductor trajectory; the RM28.5 billion figure, the AIXTRON commitment, and the broader pipeline MIDA reports — 182 potential projects worth RM14.1 billion as at 5 May 2026 (MIDA Q1 2026 Investment Performance media release) — are genuine and substantial. The more useful filter for a company evaluating Malaysia is not the aggregate investment figure but two narrower questions: does a given announcement describe front-end, equipment, or design capability (AIXTRON, the MAPC consortium’s ambition) or OSAT-layer capacity (Chipbond, MKS, and the bulk of historical Penang investment); and does the company’s own hiring plan account for a labour market where the government itself describes the skills gap as a tenfold shortfall that “no policy can bridge quickly.”

Malaysia’s semiconductor ecosystem in Penang has genuine, decades-deep advantages that are not easily replicated by a newer entrant such as Vietnam or India. The New Industrial Master Plan 2030 and National Semiconductor Strategy give the government’s ambitions a documented, multi-year framework rather than an improvised one. But a company sizing a Malaysia investment on the strength of 2026’s announcement volume alone, without separately checking whether the specific segment it needs — advanced packaging, front-end fabrication, or equipment manufacturing — has the talent base to support it, is reading only half the signal MIDA’s own data provides.

Malaysia's Semiconductor Push: Reading the Investment Signals Correctly