The chip money landed. The engineers to run it did not.

Start with the part Malaysia has got right, because it is substantial. The electrical and electronics sector pulled in RM28.5 billion of approved investment in 2025. The National Semiconductor Strategy, launched in 2024, sets a target of RM500 billion in cumulative investment by 2030, backs it with RM25 billion of fiscal support, and states the ambition plainly: move up from back-end assembly and testing, where Malaysia already holds around 13% of global capacity, into IC design, advanced packaging and innovation-led manufacturing. You can watch the thesis take physical form in Penang. Chipbond opened an advanced-packaging plant worth close to $200 million. Hanic is standing up an IC-design and advanced-packaging hub and hiring across system-on-chip design, verification, physical design and design-for-test. InvestPenang's Penang Silicon Design @5km+ is trying to knit design, test and manufacturing into a single walkable geography. The investment case is not the weak point.

The weak point is arithmetic, and it comes before any policy. The government itself concedes the industry needs 50,000 skilled engineers to meet demand as it stands today. Malaysian universities produce about 5,000 engineering graduates a year. That is a tenfold gap, and it is the kind no announcement closes quickly, because a training target is a promise about a flow and the flow is precisely the bottleneck. The NSS aims to train 60,000 engineers by 2030, which is the right order of magnitude and the right instinct. It is also a number that has to be produced year after year against a starting output that is an order of magnitude short. This is the sentence that belongs on the first slide of a semiconductor pitch, and almost never appears on it.

Then there is the leak, which is arguably the harder half. It is not only that too few engineers arrive; too many of the ones Malaysia has walk out. The Malaysia Semiconductor Industry Association reckons the country loses about 15% of its talent pool every year to Singapore, Taiwan, the United States and Europe, and the skills being recruited away, IC design, wafer process engineering, advanced packaging, are the exact front-end capabilities the value-chain climb depends on. The price signal confirms the scarcity: a Hays study this year found Malaysian packages for senior chip roles have risen far enough to top equivalent pay in Japan, driven by multinationals bidding over a shallow pool rather than deepening it. A salary line that climbs because talent is scarce, not because it is more productive, is a warning rather than a win.

The shortage is also a mismatch, not just a headcount. The World Bank's Malaysia Economic Monitor finds more than a quarter of graduates struggle to land high-skilled work, not for want of a degree but because curricula have lagged where the industry actually moved. Malaysia's own STEM enrolment target of 60% has sat below 50% since 2000. So the pipeline is at once too narrow and pointed slightly in the wrong direction, which means even closing the raw number would not, on its own, deliver the specific front-end skills the fabs and design houses are hiring for.

Into this lands a change of timing that a workforce planner would not have chosen. From 1 June 2026, Malaysia raised its Employment Pass salary thresholds sharply, the top category doubling to RM20,000 and the technical band that covers manufacturing specialists moving up to a RM7,000 floor, and it now requires employers in the middle and lower bands to file formal succession plans committing to train a local replacement. The intent is coherent with the country's push to reduce structural reliance on foreign labour. But narrowing the release valve of imported specialists at the very moment firms are staffing higher-value lines tightens both taps at once, the domestic pipeline that cannot yet meet demand and the foreign one being deliberately restricted. Good policy, awkward sequencing.

And the competition is running the same sum. Vietnam has just committed about $1.08 billion to train 50,000 engineers by 2030, and is bidding for the trainers as hard as for the trainees. That is the real contest between the two economies, and it points to an uncomfortable conclusion for economic developers well beyond Southeast Asia: the marginal constraint on moving up the value chain has shifted from capital to people, and people are the one input an agency cannot manufacture on demand. Fabs and design centres locate near the engineers, not the incentives. Malaysia's willingness to name its gap out loud is itself worth something, because most places paper over it until an investor's due diligence finds it for them.