StarHub’s Successful Trial of 5G Cloud RAN Technology to Set New Benchmark for Operators in the Southeast Asian Region

Singaporean Communication Service Provider (CSP) StarHub, in collaboration with Nokia and Dell, announced in April 2025 the successful completion of the Southeast Asian region’s first 5G Cloud Radio Access Network (RAN) trial. As part of the trial, the CSP demonstrated how carrier-grade calls and cloud-native RAN functions can be supported via Nokia’s 5G Cloud RAN solution and the Red Hat OpenShift enterprise application platform, respectively. The CSP was highly optimistic about the potential of the technology, highlighting how Cloud RAN will be a key component in helping the operator both scale its network and support deployments and implementations of Artificial Intelligence (AI) applications over its network.

As covered in an earlier ABI Research report, Cloud-Native RAN: Market Developments (AN-5914), a core fundamental of the technology is shift from purpose-built RAN hardware to hardware-agnostic Cloud-native Network Functions (CNFs) orchestrated in a cloud-native way (e.g., microservices) through a centralized data center. This introduces a number of benefits that CSPs can enjoy from moving to the cloud, such as 1) increased scalability of the network with efficient resource management; 2) supporting new enterprise and/or AI applications at the network edge; 3) support for multi-tenant workloads based on different enterprise customer requirements; and 4) facilitating the provisioning of new services quickly.

Building up this cloud infrastructure, as well as RAN AI functionality, will be crucial for CSPs to play in the growing AI and Machine Learning (ML) industry in the Southeast Asian region, where total spending on AI/ML software for private sector Business-to-Business (B2B) use among the six major countries in Southeast Asia stood at US$1.94 billion in 2024 and is forecast to grow to US$3.24 billion by 2028 (at a Compound Annual Growth Rate (CAGR) of 14%).

In fact, many CSPs in the region have already recognized this opportunity and the role that CSPs can play in the AI ecosystem. Some prominent examples include Singtel, which has launched its AI Cloud Service, RE:AI, providing enterprises and organizations with affordable access to AI applications and technologies without the need for them to maintain the underlying infrastructure to support these AI workloads. Elsewhere, Lintasarta, a company owned by Indonesian CSP Indosat, also launched its AI cloud service, GPU Merdeka, which offers up to 3.35 Terabytes per Second (Tbps) bandwidth with the deployment of eight NVIDIA H100 SXM Tensor Core Graphics Processing Units (GPUs). More detailed analysis on these developments can be found in earlier ABI Insights: “Indosat’s Heavy Investment in AI Could Be the Key to Accelerating AI Adoption in Indonesia,” “CSPs in Southeast Asia Look to AI Cloud Services to Drive New Revenue Streams,” and “SEA Telcos Must Harness their Data Center Assets and Build Mobile AI as a Service Expertise to Boost Enterprise Revenues: Insights from Singtel’s RE:AI and Nxera GoToMarket Strategy.”

While 5G networks have yet to deliver on their promises of increased monetization and revenue growth, the emergence and rapid proliferation of AI technology presents CSPs with a new opportunity to capitalize on the growth in demand for reliable connectivity and computing resources. CSPs, in their unique position of traditionally providing both connectivity and data center services, are well placed to benefit from the AI adoption trend. 

However, CSPs need to be careful to avoid facing the same challenges that came with the initial launch of 5G networks. Below, ABI Research has identified some of these challenges, along with our recommendations:

• Balance Infrastructure Rollout with Ecosystem Development: While having the underlying structure to support AI applications is important, it is equally crucial to ensure that there are applications and use cases that can leverage this infrastructure. This was an initial challenge for 5G networks, where there were not many use cases that required the advanced capabilities that 5G could offer. Building up a strong network of partnerships and sandbox environments will be critical to encourage the development and adoption of telco edge AI solutions.
• Offering Seamless End-to-End (E2E) Bundled Solutions: CSPs could also aim to support AI adoption by offering seamless and easy-to-subscribe-to E2E bundled service offerings—where connectivity, computing, and management solutions are integrated—to enhance the user experience for enterprises and/or consumers. Furthermore, partnering with System Integrators (SIs) to simplify the integration of AI solutions could be another go-to-market strategy that CSPs could consider adding into their portfolio.

In summary, the AI/ML market in Southeast Asia is well positioned for growth, with CSPs in the region being well positioned to take advantage of this trend. However, beyond infrastructure development, supporting parallel growth of the AI ecosystem is crucial for overall telco business success.