At FutureNet Asia 2026, held in Singapore on 22–23 September, Kim Krogh Andersen, Group Executive – Network, Product & Technology at Telstra, delivered a keynote presentation on how the Australian operator is reinventing connectivity for the AI era.
His presentation explored why telecommunications operators need new business models, how AI is changing network requirements, and why treating the network itself as a product could create new growth opportunities.
One of the fundamental challenges facing the telecoms industry is that demand for connectivity continues to grow, but operators have struggled to capture a corresponding increase in revenue. AI could accelerate this imbalance, as networks become increasingly important to digital services, businesses and autonomous systems. Traditional measures such as download speed and capacity remain important, but AI applications will increasingly require strong uplink performance, predictable connectivity, low latency, resilience, security and continuous availability.
At the heart of Telstra's Connected Future 30 strategy is a shift from treating networks primarily as infrastructure towards managing them as programmable platforms. This means moving beyond traditional fixed connectivity services and data allowances to capabilities that customers and application developers can access, configure and consume according to their requirements.
Telstra's approach involves three important transitions: from infrastructure to platforms, from fixed services to programmable capabilities, and from traditional product lifecycles to software-based access. The objective is to make connectivity easier to discover, consume and integrate, with network capabilities designed around customer outcomes.
Network APIs are an important part of this transformation. Through initiatives such as Aduna, which aims to simplify access to network capabilities across multiple operators, application developers could increasingly incorporate network intelligence directly into their services. Applications could therefore move from being designed around network limitations towards being able to use network capabilities as an integrated part of their functionality.
Telstra is already commercialising elements of this approach through products such as Dynamic 5G, which uses network slicing to provide differentiated connectivity, and its Adaptive Networks Centre, which gives enterprise customers greater visibility and control over their network services.
Andersen also explained that AI workloads depend on connectivity between data centres, cloud infrastructure, mobile networks and end-user devices. Telstra sees its subsea cables, fibre infrastructure and mobile network as essential components of this evolving AI ecosystem. As AI applications become more distributed and autonomous, the network will need to do much more than simply transport increasing volumes of data.
A particularly interesting part of the presentation was Telstra's explanation of three stages in the evolution towards autonomous networks. The first, AI on the network, recognises that AI applications generate new traffic and place different demands on connectivity. The second, AI for the network, involves using AI to predict problems, optimise resources and improve network assurance. The third, AI as the network, represents a more fundamental transformation in which networks increasingly use intelligence to understand intent, reason about requirements and take appropriate actions.
However, Andersen also highlighted that AI exposes existing network complexity. Legacy systems, fragmented architectures, technical debt and poor data quality can prevent operators from realising its benefits. Simplification, modernisation and the retirement of outdated systems are therefore essential foundations for greater autonomy. Telstra also emphasised the importance of shared standards, common architectures, industry collaboration and trusted governance in making autonomous networks a practical reality.
The presentation provided some interesting figures illustrating how extensively Telstra is already using AI internally. Telstra reported a 275% increase in AI-token consumption, with approximately 15 billion tokens being processed every day. These workloads span customer experience, software engineering, autonomous networks and cybersecurity.
To manage this growing demand, Telstra has developed an AI Control Plane that provides enterprise-wide visibility and control over AI workloads. The platform is designed to monitor consumption and costs across providers, models and applications, allocate workloads more efficiently, and support responsible AI governance. This work forms part of Telstra's broader transformation activities, including its collaboration with Accenture. Importantly, the objective is not simply to use more AI, but to extract greater value from each model, token and workflow.
Perhaps the most interesting aspect of Telstra's strategy is its emphasis on returning to the operator's core strength: connectivity. Rather than attempting to compete directly with hyperscalers in every area of cloud computing or AI, Telstra wants to make its network more intelligent, programmable and valuable to the wider digital ecosystem.
The opportunity is significant, but the challenge is commercial as much as technological. Operators must demonstrate that customers and application providers are willing to pay for differentiated network capabilities, rather than simply expecting more performance within existing connectivity packages.
Telstra's approach provides an interesting example of how a major operator is attempting to connect network modernisation, AI adoption and new revenue opportunities into a coherent long-term strategy. Whether Network as a Product can deliver substantial new revenue remains to be seen, but the direction is clear. As Andersen emphasised in his closing slide, there is no AI without connectivity. The challenge for operators is to ensure that they capture more of the value their networks make possible.
The complete keynote presentation is embedded below:
Related Posts:
- Operator Watch Blog: Telstra’s Bold Journey to RAN Autonomy and Beyond
- Telecoms Infrastructure Blog: Metrocells to help Telstra Increase Density for Greater Capacity and Speed
- Operator Watch Blog: Telstra Conducts 50 Tests on 5G mmWave and Small Cells to Measure EME
- Connectivity Technology Blog: Telstra and Ericsson double LTE range
