A benchmark report uncovering the telcos leading AI transformation across Asia Pacific
Executive Summary
AI is no longer an optional add-on for telcos — it’s a strategic imperative. As revenue from traditional services stagnates, telecom operators across Asia Pacific are embracing AI to reinvent themselves as intelligent, data-driven, and future-ready enterprises. This report identifies the Top 10 APAC telcos that are leading the AI transformation in 2025, based on a comprehensive benchmarking of 34 operators using Twimbit’s proprietary AI Maturity Framework.
Key insights:
- SoftBank, SK Telecom, and Indosat are far ahead of the curve, scoring between 118 and 132 out of 150, and demonstrating enterprise-wide AI deployments, sovereign AI investments, and proactive governance models.
- All top 10 telcos exhibit high alignment between AI strategy and business goals, with centralized Centers of Excellence (CoEs), cross-functional AI teams, and extensive partnerships with AI infrastructure and model providers.
- AI is being applied across four critical domains: network optimisation, customer experience, AI-native infrastructure, and new digital products/services.
- Responsible AI is now a leadership marker. Telcos like Telstra, SK Telecom, and Singtel are formalising governance frameworks and aligning with global standards like ISO 42001 and the EU AI Act.
These leaders are not just using AI — they are rebuilding themselves around it.
Introduction: The Rise of the AI-Native Telco
Asia Pacific telcos face a complex future: flatlining connectivity revenue, exponential data growth, rising costs, and fierce competition from hyperscalers. AI presents a transformative opportunity to reimagine this equation, shifting telcos from infrastructure managers to digital-first technology enterprises.
Leading telcos are embedding AI into their networks, customer operations, business models, and national infrastructure roles. From Open RAN automation to multilingual GenAI agents and sovereign AI clouds, the top performers are showing what it takes to become AI-native.
Objectives and Scope of the Report
This report aims to:
- Benchmark 34 APAC telcos on AI maturity using a six-dimensional, 30-sub-component framework.
- Identify the Top 10 telcos that are scaling AI across strategy, operations, and customer experience.
- Highlight successful deployments and best practices across AI use cases.
- Provide insights for C-level leaders, AI strategists, and policy stakeholders driving telecom transformation.
Methodology: Benchmarking 34 Telcos Across APAC
Twimbit evaluated telcos across 15 APAC markets using its AI Maturity Framework, which covers:
- 6 Pillars: Strategic Alignment, People & Processes, Technology & Applications, Data Management, AI Governance & Ethics, Innovation & Experimentation
- 30 Sub-capabilities: Each scored from Level 1 (Foundational) to Level 5 (Transformational)
- Scoring: Max total = 150 points (25 per pillar)
Data Sources:
- Public reports, earnings releases, newsroom updates
- Analyst insights and verified third-party references
- Cut-off date: May 2025
Twimbit’s 6-Pillar AI Maturity Framework
| Pillar | Description |
|---|---|
| Strategic Alignment | AI vision, roadmap, executive sponsorship, KPI linkage |
| People & Processes | Talent upskilling, AI CoEs, organizational embedding |
| Technology & Applications | AI platforms, GenAI tools, proprietary ML/LLM stacks |
| Data Management | Talent upskilling, AI CoEs, organisational embedding |
| AI Governance & Ethics | Real-time pipelines, governance, and unified architecture |
| Innovation & Experimentation | PoC-to-scale engines, AI Labs, startup partnerships |
Each telco was assessed with equal weight given to all six pillars.
Top 10 APAC Telcos to Ace AI (2025)
AI Use Case Highlights by Category
📡 1. Network Optimization
- Rakuten Mobile: World’s first fully virtualised, AI-powered network (zero-touch operations).
- SK Telecom: AI Orchestrator network platform predicts faults and automates capacity allocation.
- KT: AI-Meister agents monitor and optimise 5G performance dynamically.
👥 2. Customer Experience
- Indosat: NLP agents handle queries in Bahasa, English, and local dialects — millions served.
- Singtel: GenAI-based call summarisation and automated support journeys.
- LG Uplus: ixi-O Gemini AI assistant integrated into call centres and smartphones.
🧠 3. AI Infrastructure
- SoftBank: Graphcore chips, sovereign compute, OpenAI Japan JV.
- Telstra: AI@Scale stack, 21,000+ Microsoft Copilot licenses, Intercity AI network.
- KDDI: Osaka-based AI Data Center with NVIDIA GB200 clusters.
📱 4. AI Products & Services
- KDDI: AI-powered au PAY super app — bookings, payments, recommendations.
- NTT Docomo: Tsuzumi LLM + digital twins for mobility and urban planning.
- Rakuten: AI for Business platform with RAG, job-specific prompts, 30,000 users.
Profiles of the Top 10 Telcos
🥇 SK Telecom
- $200M in SGH, $100M in Anthropic, ISO 42001 certified.
- Sapeon X330 chip developed for AI workloads.
- AI Orchestrator network platform predicts network anomalies, automates recovery.
- T.H.E. AI Principle framework for responsible AI.
🥈 SoftBank
- Built a 25.7 exaflop hybrid AI cloud with NVIDIA.
- Launched AITRAS for AI-powered 5G RAN optimization.
- Acquired Graphcore and invested in avatarin for AI CX.
- Formed SB OpenAI Japan for domestic LLMs.
- Governance through GRCC AI Working Group.
🥉 Indosat
- Sahabat-AI (70B multilingual LLM), $200M CoE with NVIDIA.
- First AI-RAN in SEA, GPU Merdeka sovereign cloud.
- Ericsson full-stack DMP; AI Cloud via Lintasarta.
- Focused on national AI talent & infrastructure development.
🇦🇺 Telstra
- $700M JV with Accenture, RE:AI strategy.
- Insights Hut, One Sentence Summary Copilots.
- Resilience-focused AI networks with Cisco.
- Risk Council on AI & Data for ethics compliance.
🇰🇷 LG Uplus
- ₩3T AI plan + hyperscale DC with AWS.
- ixi-O (Gemini-powered agent), sLLM “ixi-GEN”.
- Agent Assist scaled from 200 to 2,300 agents.
- Compliance with GDPR, PIPA, and EU AI Act readiness.
🇰🇷 KT Corporation
- ₩7T investment in AI/infra by 2027.
- Mi:dm LLM, Palantir integrations, media AI studio.
- Sovereign cloud with Microsoft and Databricks.
- Ethics program with grievance channels and education.
🇯🇵 Rakuten Mobile
- Rakuten AI 2.0 + AI for Business for employees.
- Open RAN optimisation with Intel, Cisco, and OpenAI.
- Partnered with Science Arts and zTouch for AI apps.
- AI Code of Ethics and governance structure.
🇯🇵 KDDI
- Partnered with AWS, Google, Groq, and Elyza.
- AI-enabled metaverse, e-commerce, data platforms.
- Osaka AI Data Center with 1,000-GPU stack.
- Business ethics committee for AI governance.
🇸🇬 Singtel
- RE: AI (AI+cloud platform), GPUaaS via Nxera & NVIDIA.
- Scale AI, H2O.ai, and Nscale partnerships.
- NUS/NTU AI accelerator, PDPA compliance.
- GSMA partner in the Responsible AI maturity roadmap.
🇯🇵 NTT Docomo
- Tsuzumi (energy-efficient Japanese LLM).
- 6G AI-Native Interface with Nokia & SK Telecom.
- AI retail automation and smart life services.
- Co-CAIO model + AI governance office.
Strategic Takeaways: What the Best Do Differently
- AI Starts at the Top:
Boardroom ownership, C-suite KPIs, and cross-functional alignment drive execution. - Own the Infrastructure:
Leaders don’t rent — they build sovereign data centres, GPU clusters, and LLMs. - CoEs Are Scaled, Not Siloed:
From Airtel to Indosat, CoEs now drive execution, not just pilots. - Responsible AI = Strategic Enabler:
Risk management is embedded in delivery pipelines, not added later. - Real-Time Data Is the New Stack:
Governed APIs and live pipelines power GenAI and predictive systems. - Scale > Experimentation:
Leaders track PoC-to-production conversion, not just volume of pilots.
What Will Define AI Telcos by 2030?
- Agentic AI Systems:
Enterprises will deploy self-executing AI agents across customer and network operations. - Sovereign LLM Stacks:
Telcos will invest in localised language models and training infrastructure. - Proprietary AI Chips:
Control over cost, latency, and scale will drive chip-level innovation. - Standardised AI Governance:
ISO certifications, bias audits, and explainability will become default compliance. - PoC-to-Production Velocity:
Time-to-scale will emerge as the core metric for AI ROI in telcos.
AI is not the future of telco — it is the foundation.
The Top 10 telcos in this report show how AI can move from vision to velocity — and from pilots to profit.
Let the AI era of telecom begin.
Click here for more content on AI