Hong Kong is poised to solidify its position as a regional powerhouse for artificial intelligence and digital infrastructure with the unveiling of HKT’s pioneering 3.2Tbps AI Data Centre Inter-connect (DCI) Superhighway. This landmark announcement signals a significant global shift among leading telecommunications operators, who are aggressively transforming their fibre and optical transport networks into AI-ready infrastructure. This strategic evolution is crucial for supporting the burgeoning demands of hyperscale data centres and the increasingly sophisticated distributed GPU clusters that power the artificial intelligence revolution.
The relentless acceleration of AI workloads, characterized by their immense computational requirements, is compelling telecom operators to invest heavily in ultra-low latency, high-capacity connectivity solutions. This advanced infrastructure is essential for enabling the seamless training and deployment of complex AI models across geographically dispersed data centres, a critical capability in today’s AI-driven landscape.
The Global AI Infrastructure Surge: A Telecom Operator’s Transformation
This transformation is not an isolated event but a palpable global trend. A recent in-depth analysis by Omdia highlights a significant surge in investments by telecom operators into AI infrastructure. This includes not only the development of AI-specific data centres but also the expansion of GPU-as-a-Service offerings, the establishment of robust AI cloud platforms, and the enhancement of optical transport and data centre interconnect (DCI) capabilities. Operators such as SK Telecom, SoftBank, Singtel, Ooredoo, STC, and Iliad are actively repositioning themselves, moving beyond their traditional roles as mere connectivity providers to emerge as key architects and enablers of AI infrastructure.
The financial implications of this strategic pivot are substantial. A comprehensive report by IDC forecasts an impressive Compound Annual Growth Rate (CAGR) of 31.8 percent for AI and generative AI spending within the telecommunications sector across the Europe, Middle East, and Africa (EMEA) region between 2024 and 2029. This robust growth underscores the imperative for operators to invest in AI-ready infrastructure, not only to meet the escalating demands of enterprise services but also to drive significant operational efficiencies and transformations within their own organizations.
A Wave of Innovation: Global Operators Embrace AI Connectivity
HKT’s ambitious undertaking is part of a broader wave of similar initiatives being launched by leading telecommunications players worldwide, each striving to build the foundational networks necessary for the next era of digital innovation.
In North America, Lumen Technologies has proactively expanded its AI-ready network by strategically deploying dedicated fibre routes and high-capacity wavelength services. These crucial links connect AI data centres and hyperscale cloud providers, facilitating the ultra-low-latency communication essential for the operation of large-scale GPU clusters.
Across the Pacific, NTT Communications is spearheading advancements in AI networking through its visionary Innovative Optical and Wireless Network (IOWN) initiative. By leveraging cutting-edge all-photonics technology, NTT Communications is committed to delivering connections with exceptionally low latency and high bandwidth, specifically tailored to support the demanding requirements of next-generation AI infrastructure.
European telecommunications giants are also demonstrating a strong commitment to bolstering their AI connectivity capabilities. Sparkle, the international arm of TIM Group, has significantly expanded its AI-ready international backbone. This enhancement involves integrating high-capacity optical transport solutions that directly link major data centres across Europe and the Mediterranean region, creating a robust and interconnected network for AI-driven applications.
Meanwhile, Orange Business, the enterprise arm of Orange, is actively strengthening its data centre networking portfolio. The company is introducing high-capacity Ethernet and optical services meticulously designed to cater to the specialized needs of enterprise AI applications, ensuring reliable and high-performance connectivity for businesses leveraging AI technologies.
In the dynamic Asia-Pacific region, Tata Communications has seamlessly integrated AI-ready connectivity into its comprehensive Digital Fabric platform. This innovative approach combines global networking capabilities with AI cloud infrastructure and sophisticated GPU services, providing enterprises with the essential tools to deploy AI solutions effectively across multiple geographical regions.
HKT’s Differentiated Approach: Hollow Core Fibre and AI-Native Technologies
What sets HKT’s latest initiative apart is its distinctive combination of a formidable 3.2Tbps data centre interconnect network with the integration of advanced hollow core fibre technology. This cutting-edge fibre optic solution is projected to deliver approximately 30 percent lower latency compared to conventional fibre optic networks. This reduction in latency is a critical factor for AI applications that demand near-instantaneous data transfer and processing.
Furthermore, HKT is actively embracing and supporting AI-native technologies. This includes the integration of advanced protocols such as Remote Direct Memory Access (RDMA), RDMA over Converged Ethernet (RoCEv2), Multipath Reliable Connection (MRC), and Ultra-Ethernet Transport. The adoption of these technologies is pivotal, enabling thousands of GPUs distributed across multiple data centres to operate in perfect synchronization with minimal transmission delays, a feat previously considered exceptionally challenging.
The initial phase of this groundbreaking network is slated for completion by the end of 2026. This phase will strategically link the Lok Ma Chau Loop and Tseung Kwan O, two key locations within Hong Kong. Crucially, this infrastructure will serve as the backbone for the upcoming Sandy Ridge Data Facility Cluster, a development that will further solidify Hong Kong’s ambition to become a preeminent regional hub for AI computing and advanced digital infrastructure.
The Broader Implications: Reshaping Telecom Strategies for the AI Era
The substantial and increasing investments being made by HKT and its global counterparts are undeniable indicators of how artificial intelligence is fundamentally reshaping the strategic priorities of the telecommunications industry. The focus is no longer solely on providing traditional connectivity services. Instead, operators are increasingly dedicated to building specialized AI transport networks that offer multi-terabit bandwidth, ultra-low latency, and energy-efficient connectivity.
These advanced networks are not merely an enhancement of existing infrastructure; they represent the fundamental digital foundation upon which the next generation of AI applications, sophisticated cloud computing environments, and high-performance data centres will be built. By embracing this paradigm shift, telecom operators are positioning themselves as indispensable enablers of the AI-powered future, creating the essential pathways for innovation and progress across a multitude of industries. The commitment to developing such high-capacity, low-latency networks signifies a profound understanding of the evolving digital landscape and a proactive approach to meeting the unprecedented demands of the AI revolution.
By BABURAJAN KIZHAKEDATH
