Malaysia Plans a National AI Compute Exchange to Make AI Power More Affordable

September 29, 2026

Malaysia is planning a National AI Compute Exchange that could make advanced AI computing resources easier and more affordable for local businesses, startups, researchers and government agencies. The platform is expected to launch in the first quarter of 2027.

Table of contents
Key Takeaways
  • MDEC plans to launch the National AI Compute Exchange (NACX) in Q1 2027.
  • NACX will connect users with multiple providers of AI compute, models and services.
  • The platform is aimed at SMEs, startups, researchers, enterprises, government agencies and academic institutions.
  • MDEC wants to aggregate demand so users can access more competitive pricing and simpler procurement.
  • The bigger goal is to turn Malaysia’s data-centre investments into local AI products, companies and jobs.

Malaysia Has Plenty of AI Infrastructure — But Access Is the Problem

Malaysia has already attracted major investment into data centres and cloud infrastructure.

But having physical computing capacity does not automatically mean Malaysian companies can easily use it.

For smaller businesses, AI infrastructure can still be expensive and difficult to access.

MDEC now wants to close that gap.

The proposed National AI Compute Exchange, or NACX, is designed to act as a central marketplace connecting organisations that need AI computing resources with providers that already have them.

What Will NACX Actually Do?

Instead of every company approaching individual cloud or AI providers separately, NACX is intended to bring multiple options together in one platform.

Users could potentially compare and access:

  • AI computing resources
  • Cloud infrastructure
  • AI models
  • AI services
  • Local and global providers

MDEC says the system will help simplify procurement while giving users access to more competitive pricing.

For smaller companies, this could reduce one of the biggest barriers to AI adoption: the cost of computing power.

Why AI Compute Is So Expensive

Modern AI development can require powerful GPUs and large amounts of cloud computing.

That can become expensive very quickly.

A startup may have a good AI idea but still struggle to afford enough computing resources to train, test or operate it.

The same problem affects universities and researchers.

Large companies can usually negotiate directly with cloud providers.

Smaller users often cannot.

By combining demand from multiple Malaysian organisations, NACX could potentially give those users better access to the same infrastructure.

Who Can Use It?

MDEC says NACX is intended for a wide range of users, including:

  • SMEs
  • Startups
  • Government agencies
  • GLCs
  • Enterprises
  • Universities
  • Researchers

On the supply side, the platform can also connect:

  • Hyperscalers
  • Local data centres
  • AI infrastructure providers
  • AI model companies
  • AI solution providers

That means NACX is not simply a government-owned GPU cluster.

It is intended to work more like a national marketplace for AI resources.

Malaysia Wants to Build, Not Just Host

This fits a bigger change in Malaysia’s digital strategy.

Digital Minister Gobind Singh Deo said Malaysia should not only host digital infrastructure.

The country should also build, operate and eventually own more of the technology and businesses created around it.

That distinction matters.

A data centre creates investment and infrastructure.

But AI applications, software companies, cybersecurity platforms and digital services can create more long-term value and higher-skilled jobs.

NACX is one attempt to make sure local companies can actually use the infrastructure being built in Malaysia.

Why This Matters for Malaysian SMEs

For SMEs, the opportunity is practical.

A business may want to develop:

  • AI customer service
  • Document automation
  • Predictive analytics
  • Computer vision
  • AI agents
  • Industry-specific applications

But building those systems can require computing resources beyond what a small company wants to purchase itself.

A shared marketplace could make experimentation easier.

It may allow smaller businesses to test AI ideas without making large upfront infrastructure investments.

Closing Thoughts

Malaysia has already spent years attracting cloud and data-centre investment.

The next challenge is making sure Malaysian companies can actually benefit from that infrastructure.

The National AI Compute Exchange could become an important bridge between AI infrastructure and AI adoption.

The real test will not be how many GPUs Malaysia has.

It will be how many Malaysian companies can use them to create useful products, services and jobs.

References:
  1. Business Today Malaysia — MDEC Plans AI Compute Exchange To Widen Access For Businesses

    https://www.businesstoday.com.my/2026/09/29/mdec-plans-ai-compute-exchange-to-widen-access-for-businesses/

  2. MDEC — National AI Compute Exchange (NACX)

    https://www.mdec.my/staging/nacxdev4

  3. Ministry of Digital — MDX Forum 2026

    https://www.digital.gov.my/en-GB/siaran/MDX-Forum-2026

  4. The Edge Malaysia — Powering past the data centre boom: How Malaysia must build, not just host

    https://www.klsescreener.com/v2/news/view/1799459/powering-past-the-data-centre-boom-how-malaysia-must-build-not-just-host

Frequently Asked Questions About Malaysia’s National AI Compute Exchange

  1. What is NACX?

    NACX stands for National AI Compute Exchange. It is a proposed national marketplace connecting users with AI computing resources, models and services from multiple providers.

  2. When will NACX launch?

    MDEC plans to launch the platform in Q1 2027.

  3. Who can use NACX?

    The intended users include SMEs, startups, enterprises, researchers, universities, government agencies and other organisations needing AI resources.

  4. Will the government own all the GPUs?

    Not necessarily. NACX is designed to connect available computing capacity from different providers instead of requiring the government to own all the infrastructure itself.

  5. Why is this important for Malaysia?

    It could make AI resources more affordable and accessible, helping Malaysian companies move from simply adopting AI tools to actually building their own AI products and services.