By Hema Kadia, Founder, TeckNexus.

Every keynote at this year’s FutureNet Asia will use the word “autonomous.” Networks that heal themselves. Agents that act without waiting for a human. The AI-native telco, re-architected from the ground up. It’s a compelling story, and a lot of it is true — Asia-Pacific operators are, by the numbers, ahead of the rest of the world on this. But “ahead” and “there” are different claims, and the gap between them is where most of the interesting detail actually lives.

TeckNexus Intelligence built the AI Agent Tracker to close that gap — independently verifying AI agent deployments across 50 global telecom operators against named, on-the-record sources, and scoring each one on how autonomous it really is and how strong the evidence behind the claim actually is. No source link, no score. Here’s what that discipline found when we looked specifically at Asia-Pacific.

APAC leads the world on autonomy — and the bench is deep, not one flagship program

60% of the APAC operators we track — 9 of 15 — have reached the industry’s most advanced verified autonomy tier, the highest share of any region we measure. Europe, spread across a far larger footprint of 16 markets, sits at 38%.

Rank APAC’s 15 tracked operators on how many distinct parts of the business their AI agents actually touch — customer service, network operations, RAN, back-office — and four operators tie for the regional lead: China Mobile, NTT DOCOMO, Rakuten, and SoftBank. A close second tier follows at SK Telecom and Telstra, then LG Uplus, Bharti Airtel, Reliance Jio, and Optus form a solid middle of the field. This is a genuinely deep bench, not a single standout carrying the regional story.

What APAC’s lead is actually made of

Here’s where it gets more interesting than “APAC is ahead.” Break that lead down by capability and compare it to the global average, and APAC isn’t actually leading on the category everyone assumes — customer experience, where the region runs roughly even with the rest of the world.

The real gap is everywhere else: APAC pulls ahead of the global average on network automation, on RAN and radio-side AI, and on back-office/OSS. APAC’s advantage isn’t chatbots — it’s operational depth in the categories that don’t make headlines.

That shows up in strategy, too. Of the 15 operators we track in the region, 7 lead with a consumer-facing AI strategy, but 5 lead with network operations instead — a far larger network-first cohort than most other regions carry — and only 2 run a genuinely mixed strategy across both.

But the ceiling hasn’t moved — anywhere

Here’s the finding that matters most for anyone in the room framing autonomy as a near-term inevitability: across all 50 operators we track globally, not one has reached full, unsupervised autonomy. Zero sit at the top two tiers of the industry’s autonomy scale. The most advanced deployments — including the APAC leaders above — still keep a person positioned to review exceptions before they execute.

That’s not a knock on the technology. It’s a deliberate posture. When an AI agent gets something wrong in this industry, it’s a customer’s bill, a family’s connectivity, or a compliance obligation — and operators, almost universally, are choosing not to remove human oversight yet, even where the technology could plausibly support it. Worth sitting with, given how much of this week’s agenda is framed around the journey to L4 and L5.

Breadth and autonomy aren’t the same axis

Here’s a pattern worth watching for in your own competitive reads: the operator with the broadest program isn’t always the one furthest along on autonomy, and vice versa.

NTT DOCOMO ties for the highest breadth score of any operator we track in the region, backed by the strongest evidence rating on record — and still sits one tier below the region’s most advanced autonomy bracket.

Singtel sits at the opposite extreme: the lowest breadth score among the 15 APAC operators we track, and it still qualifies for the leading tier. A narrow program can be a genuinely autonomous one.

A broad program can still be conservative about how much it lets agents do unsupervised. Benchmarking a competitor on a single number risks missing which number they’re actually strong on.

See the full 50-operator autonomy tier and evidence strength view → tecknexus.com/intelligence/ai-agent-for-telecom-operators

The partner landscape looks different once you zoom into the region

Globally, Google Cloud is the single most-named AI partner across all 50 operators’ programs. Zoom into Asia-Pacific specifically, and the picture is far more fragmented: five players are tied for the most-cited partner in the region — Huawei, Meta, Google, LG AI Research, and Sierra — each named in three of the 53 individually verified APAC AI initiatives we track. Google Cloud, the clear global leader, shows up in just one. For anyone weighing a build-vs-partner decision this week, the message is that APAC’s AI partner ecosystem isn’t following the global pattern — it’s considerably more open, and more regionally distributed, than the worldwide numbers suggest. 

This is a small sample of a continuously maintained, 50-operator research program. The Operator Lane gives full access to the complete analysis — every operator, every region, every category, scored against the same evidence standard: see the full findings at tecknexus.com →

Hema Kadia is the founder of TeckNexus, an independent intelligence tools, research and media platform providing evidence-based benchmarking of agentic AI deployment across the global telecom industry.