Malaysia's Data Centre Boom: The AI Infrastructure Careers Powering Johor and KL

When people picture Malaysia's AI boom, they imagine model-builders in KL offices. But the country's most visible AI transformation is happening in windowless buildings on former plantation land: data centres. Johor is now Southeast Asia's largest data centre hub, hyperscalers have committed tens of billions of ringgit to Malaysian soil, and the government has staked its AI strategy on national compute. Behind every Malaysian chatbot, every sovereign cloud, every GPU cluster, sits a workforce that barely existed five years ago.

This article maps that workforce: the roles Malaysia's data centre boom is actually hiring for in 2026, what they pay in MYR, and how to get in before the biggest construction wave peaks.

Why Malaysia became Southeast Asia's compute hub

Geography did most of the work. Singapore ran out of land and power for hyperscale facilities, and Malaysia — Johor in particular — sits directly across the Causeway with cheap land, competitive electricity, and no history of moratoriums. The Johor-Singapore Special Economic Zone (JS-SEZ), launched in January 2025, formalised the spillover.

The investment numbers are staggering. MIDA-reported approved digital investments since 2021 exceed RM278 billion, with roughly RM184.7 billion tied to data centre projects. ByteDance's commitment in Sedenak Tech Park has grown past RM29.5 billion. Google built its first Malaysian data centre at Elmina Business Park in Selangor and opened a cloud region. Microsoft committed around RM10 billion across Malaysia West and Johor Bahru, AWS built its RM27 billion Cyberjaya region with three availability zones, and Oracle announced the largest single technology investment in the country's history. Johor alone now counts 15 operational data centres, 11 under construction and 25 approved projects, with pipeline capacity reaching 5.3 GW — up from roughly 10 MW of capacity in 2021.

The government has doubled down. Budget 2026 allocated about RM2 billion for a sovereign AI cloud, alongside RM5.9 billion for AI-related research and development. And in February 2026 the Prime Minister confirmed that new data centre applications unrelated to AI would no longer be approved — every new facility must justify itself as AI infrastructure. Malaysia is no longer hosting generic server farms; it is building an AI compute industry, and it needs people to run it.

The jobs the boom actually creates

This is where the popular story gets honest. Construction employs thousands of people temporarily; operations employ far fewer. A hyperscale facility may need only a few hundred staff once live. The Asia-Pacific Data Center Association projects the wider Malaysian AI industry will create 30,900 jobs annually by 2030, a large share in high-value roles: network engineers, cloud infrastructure specialists, data specialists and ICT security professionals. These are the roles worth building a career around — and they are already scarce. Jobstreet listed more than 2,200 data centre engineer vacancies in Malaysia in August 2026, many explicitly for AI and HPC infrastructure supporting NVIDIA GB200 and GB300 GPU clusters.

The roles in demand

Critical facilities engineers run the building itself — power distribution, UPS systems, cooling, fire suppression, and the metering that keeps energy costs down. In AI facilities, where a single rack can draw more power than a Malaysian kampung house, thermal management has become an engineering discipline of its own. This role suits electrical and mechanical engineers more than computer scientists.

Data centre operations technicians are the on-the-ground staff who rack servers, run cabling, monitor alarms and handle break-fix work. It is the most common entry point, and it is shift work — but operators consistently promote technicians who learn the automation stack.

Network engineers carry the real prestige. A GPU cluster is useless if data cannot move through it, and AI training runs demand high-speed fabric design — spine-and-leaf topologies, BGP routing, DCI links between campuses — that traditional enterprise networking rarely touches. Network specialists with AI-fabric experience are among the hardest hires in the country.

AI infrastructure engineers (sometimes called GPU platform engineers) sit between the cluster and the models. They install and manage GPU servers, orchestrate workloads with Kubernetes and Slurm, tune drivers and firmware, and keep utilisation high enough that the economics work. This is the role where software engineers can cross over.

Cloud infrastructure architects design the multi-tenant platforms on top — the Kubernetes platforms, identity and access layers, and cost controls that let a Malaysian company rent AI capacity from a sovereign cloud. Security and compliance specialists round out the list: data centres are audited against ISO 27001, SOC 2 and increasingly the Personal Data Protection Act, and every hyperscaler needs local people who understand both the frameworks and the regulators.

What it pays in MYR

Salaries vary sharply between Johor and Klang Valley, and between operators and hyperscalers. Typical 2026 reports, cross-referenced from Indeed and Glassdoor, put data centre technicians around RM3,700 to RM4,000 per month at entry level. Data centre engineers average roughly RM4,900 nationally, with Johor Bahru averages nearer RM7,300 and experienced engineers passing RM11,000. Critical facilities engineers in specialised facilities are reported well above that — salary trackers cite annualised figures near RM139,000, roughly RM11,600 per month. Team leads and managers run RM15,000 to RM18,000, and senior specialists in GPU infrastructure, AI networking or site leadership are being quoted RM25,000 to RM30,000 by the most aggressive operators.

The pattern is simple: generic facilities work pays like IT operations; AI-specific infrastructure work pays like software engineering. A technician who learns GPU cluster administration can out-earn a data scientist within a few years — supply and demand are that skewed.

Skills that get you hired

Malaysian operators are hiring for a specific skill stack. On the facilities side: critical environment experience, UPS and diesel generator knowledge, cooling systems (increasingly liquid cooling for AI racks), and energy efficiency measurement. On the IT side: Linux administration, Kubernetes, automation with Ansible or Terraform, and networking fundamentals up to BGP. On the AI side: GPU cluster basics — knowing what CUDA, NCCL and Slurm do, and understanding why an all-to-all training job is different from a web workload — is now a genuine differentiator on Malaysian CVs.

Certifications still carry weight in this industry precisely because it is young: CDCP and CDCS for data centre design and operations, CCNA or CCNP for networking, CKA for Kubernetes, and NVIDIA's growing certification family for GPU infrastructure. Entry-level candidates without degrees are hired on the facilities track; the engineering tracks increasingly demand demonstrable Linux and networking projects, which is good news for career-changers who have been building homelabs and contributing to open source.

Where to start, and the honest caveats

The jobs cluster in three geographies: Cyberjaya and the Klang Valley, where AWS, EdgeConneX and Google anchor enterprise-scale facilities; Selangor's Elmina corridor around Google's campus; and Johor's Sedenak, Iskandar Puteri and Kulai, where ByteDance, AirTrunk, Princeton Digital Group and a dozen other operators are building fastest. Johor pays a premium because demand is outrunning the local talent pool — recruiters report engineers commuting from KL or relocating from Penang's E&E sector.

The caveats are real and worth weighing. Data centres create fewer permanent jobs per ringgit invested than semiconductor fabs, which is precisely why the government is pairing the boom with sovereign AI ambitions rather than leaving it as pure real estate. Energy and water strain is mounting — communities in Gelang Patah protested their first data centre project in February 2026 over dust and water concerns, and the freeze on non-AI applications is partly a response to that pressure. And there is a genuine localisation question: historically, a share of specialised engineering roles went to expatriates. That is closing as local training programmes scale, but it means Malaysian candidates need actual skills, not proximity.

The takeaway

Malaysia is building compute infrastructure at a pace Southeast Asia has never seen, and the AI-focused freeze on new applications guarantees the facilities that do get built will need AI-savvy operators, not just server farms. The window for entering this field is now: the construction wave peaks before the operations wave, and the people hired in 2026 and 2027 will become the senior engineers and site leads of the 2030s. If you can pair a traditional infrastructure skill — power, networking, Linux — with even basic GPU fluency, you are targeting one of the most supply-starved, best-paying corners of Malaysia's AI job market.