AI and Data Centres Are Hiring in Malaysia — Mostly Not for the Jobs You Think

When students hear "AI jobs", they picture someone training models. Almost none of Malaysia's actual AI-driven hiring looks like that.

The country's data centre build-out — the physical infrastructure AI runs on — is recruiting electrical, mechanical, automation and facilities professionals, and competing with semiconductor manufacturers for exactly the same technical talent. Industry bodies project AI will create in the order of 30,900 jobs a year by 2030, with a skilled-labour shortage already visible.

That gap is the opportunity, and it is not where students are looking.

The four tiers of AI-era hiring

1. Build and run the buildings. Electrical distribution, cooling and HVAC, generators, fire suppression, building automation, security. Entry: SKM certificate or diploma in electrical, mechanical or air-conditioning trades. This tier is hiring hardest and asks least at entry.

2. Run the machines. Data centre technicians, network operations, hardware installation and replacement, monitoring. Entry: diploma in IT, networking or electronics, plus vendor certifications.

3. Build the software. Cloud engineering, data engineering, MLOps, security. Entry: degree in computing or a strong portfolio plus certifications.

4. Build the models. Research and applied ML. Entry: degree plus postgraduate study, and far fewer positions than the public imagines.

Most careers advice addresses tier 4. Most of the hiring is in tiers 1 and 2.

Why tier 1 deserves a serious look

A data hall is, physically, an enormous electrical and cooling problem. Every megawatt of it needs people who can install, maintain and troubleshoot high-load electrical systems and industrial cooling — and those people are trained through TVET, not university.

That means an SPM leaver can enter a growing, capital-intensive sector in one to two years through an SKM programme financed by PTPK, rather than four to five years through a degree. See the PTPK skills loan guide for the financing and the TVET and SKM guide for how the levels work.

What actually makes you employable in tiers 2 and 3

Not the degree name. Three things, in order:

1. A working portfolio. For software, something deployed that other people use. For infrastructure, hands-on time with real equipment — a home lab counts.
2. Vendor certifications. Cloud, networking and security certifications are read by employers as a concrete claim, unlike a course completion badge.
3. Evidence you can operate in a team. Almost all of this work is shift-based, procedure-heavy and collaborative.

Two honest cautions

"AI" on a course brochure is not a qualification. Many programmes have been rebranded rather than redesigned. Ask what the graduates of the last two cohorts are actually doing, and check the programme's accreditation on the MQA register — see how to check accreditation.

Policy can move fast. Malaysia's data centre policy has been tightened to prioritise AI-grade projects over general capacity, which changes what gets built and where. Sector projections are estimates, and a plan should have a second use for the same skills — electrical and cooling skills transfer to manufacturing, healthcare facilities and construction.

Where to start reading

Look at the specific occupations, their entry requirements and typical pay bands on the careers pages, and see which programmes actually lead to them in the programmes catalogue.

Job-creation projections are industry estimates reported in 2026, not government guarantees. Treat them as demand signals for planning, not as promises of employment.