More power, more renewables, more complexity: Malaysia’s next energy challenge
Why utilities in Southeast Asia need to connect data, AI and resilient infrastructure to keep power flowing reliably, affordably and securely
- By:
- Henrick Choo
Malaysia’s growing economy needs more electricity — for data centers, smart manufacturing, electric vehicles and wider electrification. The country also needs to integrate more renewable energy while keeping the grid stable — affordably and securely.
Southeast Asia is emerging as one of the world’s major centers of energy-demand growth. The International Energy Agency (IEA) says the region accounts for nearly 20% of global energy demand to 2035 under today’s policy settings, even though it accounts for only 9% of the world’s population. Electricity demand is already growing twice as fast as overall energy use.
The pressure is especially visible in Malaysia. The country’s Energy Commission has identified industrial expansion, data centers, electrification and new technologies as drivers of rising demand and argues that the cost, reliability and carbon profile of electricity will increasingly influence investment and economic competitiveness.
For major utilities, generating enough electricity and moving it reliably from power stations to customers is now only the starting point. They also need to know and understand the status of millions of connected assets, anticipate what will happen next and make better decisions quickly — sometimes in mere seconds.
The energy transition now depends as much on better data, faster decisions and resilient systems as it does on engineering.
A more complicated grid needs better infrastructure
Conventional grids were designed largely to accommodate predictable, centralized generation and one-way flows of electricity. When you add utility-scale solar, rooftop photovoltaic systems, batteries, electric vehicles and customers that both consume and generate electricity, those flows become much more dynamic.
In Malaysia, Tenaga Nasional Berhad (TNB), the primary national power utility, acknowledges this transition from one-way electricity distribution to a decentralized system.
However, this will require substantial investment in physical infrastructure. In Southeast Asia, the IEA estimates that transmission and distribution networks will need to more than double in length by 2050, while investment in grids and storage needs to rise from about $13 billion today to $50 billion in 2050 if existing climate pledges are to be met.
A layer of intelligence completes the picture
Building more infrastructure is only half the story. The other half is getting more intelligence from it.
Traditionally, the condition of a transformer serving a fast-growing industrial area might be assessed through scheduled inspections, maintenance records and alarms triggered when operating thresholds are crossed. Now, you can add weather data, load patterns, thermal readings, maintenance history and real-time operating data, and AI can spot subtle changes in those signals long before a conventional alarm sounds.
Utilities can also move beyond fixed maintenance schedules and prioritize assets that need attention by understanding what’s changing, why, and the likely consequences of leaving an issue or alert unattended.
On large networks, that translates into fewer unplanned outages and more targeted maintenance spending. NTT DATA already applies AI, IoT and advanced analytics to areas such as asset performance and predictive maintenance in energy and natural-resources environments.
For national utilities operating thousands of miles of distribution lines and large numbers of substations, even incremental improvements can make a huge difference.
Collecting data isn’t the same as acting on it
Utilities aren’t short of data, or the devices used to collect it. Meters, substations, supervisory control and data acquisition (SCADA) systems, weather services, enterprise platforms, field crews, drones, satellites and connected devices all contribute information.
When an electricity distributor sees an unusual increase in losses in one part of the network, analyzing meter data, network topology, historical consumption, asset data and field information together makes it much easier to narrow down the possible causes with the help of AI.
The same principle applies to resilience. Combine geospatial information with vegetation, weather, flood, asset-condition and network data, and teams can build a much richer picture of where infrastructure is exposed and where to intervene first — an increasingly important capability in a changing climate.
However, consolidating so many types of data into an operationally useful foundation is complex. AI cannot compensate for fragmented, poorly governed or inaccessible operational data. So, utilities have to connect IT and OT data while maintaining each dataset’s context, including which asset generated it, where that asset is in the network and what normal behavior looks like.
Once that data foundation is in place, the next question is where and how quickly the resulting intelligence should be applied.
Deploy intelligence where the work happens
Not every operational decision needs to involve sending all the relevant data to a distant cloud. When cameras are inspecting equipment or drones are surveying transmission corridors, there may be good reasons to process that information locally. Connectivity may be intermittent, or moving large volumes of raw data may be expensive or present a security risk.
This makes edge and physical AI especially relevant to utilities. Edge computing processes data close to substations and field equipment, enabling AI to identify anomalies and trigger workflows without sending raw data elsewhere. NTT DATA’s Edge AI, for example, discovers, contextualizes and normalizes data from OT devices so that it can be used by operational AI applications.
Connectivity is the other part of the equation. Private 5G provides predictable, low-latency communications in operational environments, with traffic segmentation helping prioritize critical workloads such as SCADA communications. Satellite connectivity and geospatial intelligence can extend the picture in large or difficult-to-reach areas.
The point isn’t to put AI, private 5G or IoT everywhere but to identify operational problems where faster, richer information changes the outcome, then build the architecture around those problems.
AI can reduce energy waste while creating demand of its own
There’s something else utilities cannot ignore: AI is both part of the solution and part of the demand problem. Malaysia’s rapid growth in cloud and AI data centers is introducing new considerations for electricity demand, infrastructure planning and regulation, according to the Energy Commission.
Utilities can use AI to forecast demand, optimize asset use, integrate renewable generation and improve energy efficiency — but the infrastructure running AI consumes electricity too.
NTT DATA research estimates that a single LLM query can consume 0.3Wh to 1.0Wh of electricity; at millions of queries a day, inference becomes a significant load on the grid.
So, pursuing the goal of “sustainable AI” comes down to practical choices about models, infrastructure, workload location and energy use. For example, does a particular task require a large model running in a data center or could a smaller model at the edge perform it more efficiently? Those choices can have a substantial effect: our research estimates that techniques such as model quantization can reduce inference energy use by 50% to 70% with minimal impact on accuracy.
Sustainability has to be designed into AI from the beginning, not measured only after deployment. That’s a key principle for utilities: The technology used to optimize the energy system should not undermine the outcomes it is intended to improve.
When everything connects, resilience becomes the common denominator
There is an unavoidable consequence of connecting more assets, data and intelligence: The attack surface grows. Smart grids now link systems that used to be relatively isolated, sensors are connecting to networks, AI is consuming data from multiple sources, and IT and OT environments are exchanging information.
Cybersecurity cannot be separated from operational resilience. While a compromised enterprise application may disrupt business processes, a breach affecting operational technology can disrupt electricity supply and, in some cases, create safety risks.
The answer isn’t to keep IT and OT apart, because much of the value of digital operations depends on that interaction. Instead, utilities need to know what’s connected, control how systems communicate, manage identities and access, detect anomalous behavior and design environments that can recover when something goes wrong.
NTT DATA supports OT security in areas such as industrial control systems, including SCADA monitoring, as well as access control, encryption, network assessment and continuous risk management.
AI adds another governance layer. As models move from providing information to recommending or initiating actions, utilities need clarity about who can approve an AI use case, who owns the resulting decision and who remains accountable once the system is live. Those questions are also becoming part of Malaysia’s wider conversation on AI governance.
The opportunity is in connecting the pieces
The energy transition, grid modernization, AI, sustainability and cybersecurity cannot be treated as separate transformation programs. They depend on one another.
More renewable generation requires a more flexible grid — which, in turn, requires better visibility. Better visibility depends on connected assets and trusted data. But visibility alone isn’t the goal. The value comes from turning that intelligence into decisions and actions that improve how the grid operates. That creates opportunities for AI and analytics, and moving intelligence closer to physical operations requires edge computing and dependable connectivity. And connecting all of these environments makes cybersecurity, governance and resilience more important.
Our role is to help utilities connect those pieces and turn better intelligence into measurable operational impact, drawing on four decades of experience in Malaysia as well as our regional and global capabilities. We bring together data and AI, networks and edge infrastructure, IT and OT, cybersecurity and sustainability to achieve specific operational outcomes, such as accurately predicting asset failure or integrating more distributed renewable generation without compromising grid stability.
These solutions lead to measurable improvements in how utilities operate, and they will become increasingly important in Malaysia and other Southeast Asian countries that need electricity infrastructure which can respond more quickly to new risks and fluctuating demand.
For utilities building a future in the region, the opportunity is to turn a growing wealth of operational intelligence into an energy system that can respond more quickly and efficiently to whatever comes next.
What to do next
Author
Henrick Choo
Henrick Choo is Managing Director, Malaysia, at NTT DATA, Inc. With over 30 years of information and communications technology experience, he assists clients in their digital transformation journey and has established close relationships with global technology partners within the industry.