Will 2026 bring more AI-driven growth for telcos in Asia-Pacific?

If 2025 was the year where organizations focused on experimenting with AI, 2026 is shaping up to be the year that enterprises across Asia Pacific move from experimenting with AI to full-scale production and deployment. However, these same organizations must do so with the realization that network will be the bottleneck that determines who succeeds.

For APAC operators, the opportunity isn't just "more data", but instead about addressing the commercial and technical disconnects that still plague the region. While our recent Enterprise Horizons findings found that 82% of organizations in Asia are optimistic about growth in the coming year , both geopolitical and macroeconomic uncertainties are forcing a reassessment of the infrastructure that enables AI to work its magic. Here are the five realities for the year ahead, as drawn from our conversations from the floor of the recent ITW Asia 2025.

The "spiky" reality of AI workloads

It is no longer just about having big pipes at the ready; the new imperative is managing unpredictability instead. Standard steady-state capacity planning as we know it, is dead.

Why it matters: AI training and inferencing create massive, irregular surges, or what experts call "spiky" traffic. For example, events like Black Friday or 11.11 sales in Asia can see demand spike exponentially before vanishing. Consequently,

telcos must stop selling static fat pipes and start selling elasticity.

According to IDC, 94% of companies in Asia report that their networks currently limit their ability to run large data and AI projects. The second biggest inhibitor is the inability of networks to scale up and down flexibly on demand, a challenge cited by 41% of Asian technology leaders. Having the ability to scale up for an hour of training and scale down immediately will be organizations’ new baseline for agility, as well as an economic necessity.

The rise of the hybrid sovereign network

In APAC, data sovereignty is fragmenting the cloud map. We are seeing a blurring of lines between global public clouds and local "sovereign" clouds due to national AI security concerns. This aligns with broader market sentiment, where geopolitical turmoil is ranked as a top risk factor for growth for one in five of Asian technology leaders.

Enterprises don't just want a connection to a global hyperscaler; they need secure, low-latency routes to local GPU clouds within their national borders. In 2026, the hybrid connectivity model will come out on top, with telcos that can weave together terrestrial fiber, subsea cables, and satellite into a single, compliant fabric, likely to dominate.

At the same time, the dual factors of sovereignty and security will be the new "Latency”. One in five Asian business leaders are looking to greater investments in matters of governance, risk, and compliance as a business priority. In addition, with cybersecurity remaining a top business priority , customers will pay a premium for networks that keep their AI training data within specific jurisdictional boundaries.

The NaaS reality check: Addressing the "last mile"

Network-as-a-Service (NaaS) is great but is meeting challenges at the edge in ASEAN. You can automate the core network in seconds, but if the local loop (the last mile) requires a manual truck roll or costs, say 70%, of the total circuit, the "On-Demand" promise becomes difficult to fulfil.

Our data shows similar friction; nearly one in five of Asian technology leaders struggle with the ability to source and manage multiple connectivity providers. When sourcing networks outside their home regions, organizations in Asia also face significant network reliability (39%) and provider selection issues (38%).

In 2026, the industry must solve inter-carrier orchestration to automate this fragmentation. The winners won't be those with the best APIs, but those who solve the commercial aggregation of the fragmented APAC last mile.

Talking operational AI from reacting to pre-empting

Telcos are finally drinking their own champagne and moving from using AI for baseline use cases like customer service chatbots, to more strategic applications such as for predictive maintenance. This shift is essential as efficiency remains a top three one business priority for organizations in 2025 , driven by the need to streamline operations and reduce waste.

Operators are already seeing massive efficiency gains; a speaker at ITW Asia noted being able to reduce tickets coming in from 100 down to 30 by filtering noise and predicting failures before they happen.

In 2026, we will see greater adoption of the Digital Twin model. We will see operators building full digital replicas (Digital Twins) to serve as 'sandbox environments.' This allows operators to simulate failures and test changes safely, potentially reducing POC cycles from a year to just months. Furthermore, it enables the prediction of outages (like subsea cable faults) based on "drift" data days in advance. This directly improves the bottom line, by driving down OpEx by predicting fiber cuts and power outages before they occur.

Intent-based networking

A final trend for 2026 is that we will talk less about technology (bandwidth, jitter, latency) and start talking more about the outcomes that matter. Here is a simple analogy: An enterprise Data Scientist doesn't care about packet loss; they care about model accuracy.

As such, in 2026 we will see the emergence of Intent-Based Networking. A customer simply states their intent (e.g., "I need to run this AI inference model for 100 branches"), and the network automatically configures the latency, routing, and redundancy to deliver that business outcome.

This is crucial because organizations are currently facing a massive skills gap; 38% of Asian organizations  struggle to find or retain networking skills. By automating the "how" and focusing on the "what," telcos can bridge this gap. Instead of selling dumb pipes, telcos can shift to selling AI performance.

The shift from building AI to monetizing AI

2026 looks to be an exciting year as the industry pivots from "Building AI" to "Monetizing AI." However, for APAC decision-makers, it is also the year to take a step back and audit your infrastructure.

If your network requires a phone call to upgrade capacity, or if your "automation" stops at the data center door, you may be at risk of missing the AI growth wave. The future is spiky, sovereign, and intent based. Are you ready?

Connect with us to see how we can support your AI goals.