What AI Means for Telecom

At two recent conferences, DTW Ignite and MWC Shanghai, the role of AI emerged as the dominant theme. But there remains tremendous debate about the business opportunity for AI among telecom operators and suppliers.

Should telecom operators become AI factories and charge for tokens, or should they focus on charging for data as now? How can telecom build the right digital infrastructure for an AI world? Will the rise of robotics lead to opportunities for operators in physical AI at the network edge? What impact will AI have on network requirements? How will users’ AI agents and devices evolve? What role will the hyperscale AI providers have in telecom? One area that there’s clear consensus on is using AI to improve operations and maintenance as networks become more autonomous.

South-East Asia Is a Crucible for Global Competition

The next major occasion for these topics to be discussed will be in Malaysia in early September. The GSMA’s M360 Asean conference and ZTE’s 15th Global Summit and User Congress will be co-located. They take place from 8 to 10 September in Kuala Lumpur.

From a telecom perspective, this part of the world is striking because all global suppliers are competing against each other. Unlike other regions where either Western or Chinese suppliers are at a geopolitical disadvantage, here everyone operates. The Asean region is also dynamic with an optimistic outlook, a fast-growing urban population and vibrant cities embracing new devices, mobile-first lifestyles, high-speed 5G and fibre broadband.

Malaysia also has an innovative and unusual 5G network landscape — even compared with other parts of south-east Asia. Initially the Digital Nasional Berhad (DNB) ran a single 5G wholesale network. Each existing mobile operator offered service on it using a multi-operator core network (MOCN) setup. However, this changed with the launch in August 2025 of a second 5G network operated by U Mobile and branded as Ultra5G. ZTE and Huawei are suppliers for Ultra5G and Ericsson is the supplier for DNB.

As a new network, U Mobile opted to start with a 5G standalone service. With DNB’s offering also delivering 5G standalone as well as 5G-Advanced features, Malaysia is well-placed to support AI services, robotics and the latest in 5G-enabled innovations.

Recent announcements in Malaysia highlight these dynamics. ZTE signed a memorandum of understanding with U Mobile to pilot AI features on 5G using ZTE’s Air Core platform. Ericsson announced new security to spot rogue transmitters on DNB’s national 5G network, which Ericsson operates. Mavenir used its OSS/BSS platform to support a re-invention of mobile virtual network operator Tune Talk.

The Rise of AI Infrastructure in the Digital Economy

The role of the public sector and the telecom sector is a topic of debate in South-East Asia. Economies throughout the region are seeking to grow and become major global players. Like everywhere, AI is at the forefront of conversations because of the enormous investments going into AI data centres, often called AI factories, and the associated infrastructure for connectivity and power generation that is needed to support them.

Small and medium businesses in Malaysia are optimistic about AI: 41% see it as enabling new products and services, and 36% are piloting or experimenting with AI, according to a survey supported by Red Hat and Malaysia’s National AI Office, published in October 2025. The drivers are cost savings, mentioned by 54% of respondents, and flexibility, which 51% cited.

AI is central to Malaysia’s planned economic and social development in the Thirteenth Malaysia Plan, covering 2026 to 2030. The country intends to become a regional leader in AI, digital technology and renewable energy. Also, it aims to create 500,000 new jobs in the digital economy. This plan ties into themes shaping the events in early September about the rise of intelligent infrastructure and creating value from digital transformation.

At DTW Ignite, AI was everywhere but, because of that event’s heritage, a lot of the AI demonstrations and announcements centred on using AI to automate the network and manage rising complexity. The latest developments have shifted to AI agents and the idea of inter-agent interaction. Agents should also be able to streamline operations and maintenance and accelerate time to fix issues. But AI is much more important for telecom networks than the uses discussed at DTW, and we’ll see the wider AI opportunity front and centre at M360 Asean and ZTE Summit in Malaysia.

Telecom Has an Extensive Role in AI

Networks play an important part in AI. There’s a shift to distributed AI with models and agents living on users’ devices, as well as throughout the network as part of an AI grid, and in large cloud data centres. Each AI instance will have different access to information. Those on devices will be able to act on personal or private information. In the cloud, AI models can be larger and draw on the enormous range of ever-changing information about the world. Similarly, in companies, AI will have privileged access to confidential information such as customer lists, pricing, suppliers or stock levels. Telecom networks are essential for connecting AI agents together and enabling these models to interact.

Network equipment suppliers have different strengths in this new AI ecosystem. Some are specialists in wireless networks alone. Others have extensive optical, fixed access and data centre product lines. ZTE is one of these diversified AI providers that has an extensive product line that includes rack-mounted storage, general purpose servers and GPU units for use in the data centre.

The base station is emerging as a particular new focal point for telecom AI. To improve radio network efficiency, numerous companies are now looking to use machine learning algorithms to boost spectral efficiency and shift to a quicker, software-based innovation cycle. Also, this is a part of the vision of 6G’s designers to make 6G “AI-native”.

The idea of the AI-RAN goes beyond using AI to improve performance. Essentially, the intent is for base stations to become mini-AI data centres able to process workloads close to where they originate. One of the proposed uses for AI-RAN is to process sensing data where the base station’s radio becomes, in essence, a form of radar to sense weather or objects that aren’t cellular-connected. Integrated sensing and communication (ISAC) is part of the vision for 6G, but there are examples of sensing already on 5G. For example, at Dalian Changhai Airport, China Unicom has deployed a 5G mmWave-based sensing solution to help detect birds and unauthorized drones.

Robotics is another key area for AI-RAN. The concept is to have cellular-connected robots where some of the GPU-heavy processing is offloaded to the edge of the network. This helps with power consumption on the robot and reduces to complexity of the silicon needed to build into the robot. Asia is a hotbed of robotics innovation with especial emphasis on humanoid robots, as well as quadruped and wheeled robots.

Trust and Sovereign Capabilities Are a Foundation for Growth

Part of the attraction of putting mini-AI data centres into an operator’s telecom network is about trust. Telecom operators in many countries see themselves as well-placed to run sovereign technology infrastructure. This is important for AI, but also for other digital needs for the public sector, for enterprises and for the economy. Other sectors, like healthcare, media and manufacturing, can also benefit.

In Malaysia, we’ll see discussion of the role of connectivity, AI and technology to support growth in the economy. It’s an exciting time. Although it’s a long way from the traditional centres of technology innovation, this region has an extremely competitive landscape and a vast appetite to develop using technology to accelerate growth.