Education & Talent / AI news for Malaysia
APU opens Malaysia's first NVIDIA DGX Spark AI supercomputing lab
Asia Pacific University has put NVIDIA DGX Spark supercomputers into students' hands, aiming to move Malaysia from AI users to AI creators. The lab's first research and talent results are still to come.
In brief
- Asia Pacific University of Technology & Innovation launched Malaysia's first AI supercomputing lab powered by NVIDIA DGX Spark systems at its Technology Park Malaysia campus on 4 September 2026.[1]
- Each DGX Spark pairs a GB10 Grace Blackwell Superchip with 128 GB of unified memory to run and fine-tune language models of up to 200 billion parameters locally.[1][2]
- Higher Education Minister Datuk Seri Dr Zambry Abdul Kadir officiated the launch and said AI should be embedded across all fields of study, not only technical ones.[3]
A university opened a DGX Spark lab and tied it to national AI capability
Asia Pacific University of Technology & Innovation (APU) has launched what it describes as Malaysia's first AI supercomputing lab powered by NVIDIA DGX Spark systems. The lab opened at APU's Technology Park Malaysia campus on 4 September 2026 and was officiated by Higher Education Minister Datuk Seri Dr Zambry Abdul Kadir.[1][3]
Each DGX Spark unit is built around NVIDIA's GB10 Grace Blackwell Superchip with 128 GB of LPDDR5x unified memory, delivering up to one petaFLOP of AI performance. NVIDIA positions the machine as a personal AI supercomputer that can run, train and fine-tune large language models of up to 200 billion parameters without sending data to a public cloud.[1][2]
APU chief executive Datuk Parmjit Singh said the lab is aligned with Prime Minister Datuk Seri Anwar Ibrahim's call for Malaysia to move beyond being an AI user. The stated goals are widening access, building Malaysian AI talent, strengthening sovereign AI capability and moving students from using AI to creating it.[3][1]
What the lab puts in students' hands
The lab gives APU students and researchers direct access to DGX Spark hardware for coursework, experimentation and research. APU says the systems will support disciplines including computer science, artificial intelligence, quantum computing, data science and business analytics, software engineering, computer engineering and mechatronic engineering.[1]
Running models locally matters for teaching and for Malaysia's data-sovereignty goals. A student can experiment with a large open-weight model on campus hardware without exporting prompts or data to an overseas cloud service. NVIDIA's local-AI account describes the same DGX Spark family running open models such as a 30-billion-parameter Nemotron release and, across two clustered units, larger mixture-of-experts models.[2][1]
A sovereign-AI argument, not only a hardware purchase
APU's framing is explicitly about national capability. Parmjit Singh said AI literacy must go beyond using ready-made tools and that the lab is focused on strengthening sovereign AI capability. The announcement links the project to a broader government push in which Malaysia wants to build its own AI capability rather than remain a consumer.[3]
That ambition still needs measurable outcomes. The launch does not yet state how many students will use the lab each semester, how access is scheduled across APU's programmes, or which research projects will run first. DGX Spark systems are also used by institutions such as Stanford, Harvard and Arizona State University, which shows the hardware's pedigree without proving what APU's lab will produce.[1]
The minister adds a human guardrail
Minister Zambry used the launch to argue that AI should be embedded across every field of study, including the humanities, international relations and management, so curricula stay relevant. He was reported as praising APU's use of the technology in an International Relations course.[3]
He also drew a boundary. Zambry said AI should be used responsibly and not overshadow students' ability to think critically, write and develop their own ideas. That caution connects the lab to the upcoming Malaysia Higher Education Blueprint 2026-2035, where universities will be expected to adopt AI without surrendering independent reasoning.[3]
Why Malaysia should care
For Malaysia, the lab is an early test of whether giving students direct access to supercomputing hardware can turn the country's stated sovereign-AI ambition into home-grown AI builders rather than consumers of foreign models.
Students and graduates
Direct access to supercomputing hardware could improve AI coursework, but no intake or scheduling details are published yet.[1]
Practical move: Check APU's programme pages for how lab time is allocated before expecting guaranteed access.
Universities and TVET
The model shows a private university buying sovereign-AI capability outright, but the cost and operating model are not disclosed.[1][3]
Practical move: Watch whether public institutions can replicate the model within their budgets.
Malaysian AI employers
The lab aims to grow the local talent pool, but graduate numbers and job outcomes will take years to measure.[1]
Practical move: Track the lab's research output and graduate placement rather than the launch itself.
What Malaysians can do now
- Separate the lab's stated ambition from proof of student usage, scheduling and research output.
- Ask APU how lab access is shared across its seven named disciplines and whether external researchers can use it.
- Watch for the first named research projects and any published results before treating the lab as an operating programme.
What we still do not know
The launch leaves operating details open.
- How many students will use the lab each semester and how access is scheduled across programmes.
- The capital cost of the DGX Spark systems and who funds ongoing operations and maintenance.
- Which research projects, investigators and datasets will run on the hardware first.
- What measurable talent, research or sovereign-AI outcomes APU will report after the launch.
Sources
- 1.APU Launches Malaysia's First AI Supercomputing Lab Powered by NVIDIA DGX Spark for Education & Innovation Asia Pacific University of Technology & Innovation, 4 September 2026
- 2.Sparks Fly: NVIDIA Accelerates Local AI at IFA 2026 NVIDIA, 3 September 2026
- 3.Higher Education Institutions Must Embed AI Across All Fields – Zambry Bernama, 4 September 2026
- 4.NVIDIA Announces DGX Spark and DGX Station Personal AI Computers NVIDIA
