ZTE unveils AI-driven autonomous network strategy at TM Forum's DTW Ignite

• ZTE has unveiled AI-driven autonomous network strategy at Digital Transformation World (DTW) 2025, and emphasized that autonomous networks have become the cornerstone of operators' digital transformation
• ZTE launches the AIR Net Solution with a tri-engine architecture and open-sources its Co-Sight Framework, driving L4+ network autonomy through interconnected engines and innovations in data creation, advanced Graph Search, and cross-vendor interoperability

Copenhagen, Denmark, 27 June 2025 – ZTE Corporation (0763.HK / 000063.SZ), a global leading provider of integrated information and communication technology solutions, has unveiled AI-driven autonomous network strategy at Digital Transformation World (DTW) 2025 held in Copenhagen, Denmark.

Mr. Dai Yanbin, Deputy GM of Service & Data Intelligence Product Line at ZTE, delivered a keynote titled "AI & Data Solid the Path to L4 Autonomous Network", during the session, emphasized that Autonomous Networks have become the cornerstone of operators' digital transformation. "This is no longer a future vision – it's today's imperative," he noted, pointing to China Telecom's well-defined roadmap for achieving L4 high-level autonomy at the joint presentation session from China Telecom and ZTE.

Dai Yanbin delivered a keynote speech at Digital Transformation World (DTW) 2025

The AI Revolution: From GPT to DeepSeek

The presentation traced AI's exponential growth since the "GPT Moment" of 2022 through the 2025 "DeepSeek Moment". Stanford's 2025 Artificial Intelligence Index Report confirms that AI now surpasses human baselines across all 8 benchmark categories – from image classification to PhD-level scientific reasoning.

"While AGI maturation continues," Mr.Dai observed, "vertical industries like telecom are already primed for widespread generative AI deployment in high value scenarios." With inference ability being better than human baseline and costs plummeting by 100-1000x in some scenarios, AI has become the decisive battleground for operator competitiveness.

ZTE's AIR Net: The Evolution towards L4+ Network Autonomy with Tri-Engine Architecture

ZTE's response is the AIR Net Solution, built on three interconnected engines: AI Large Language Model Engine,Big Data Engine, and the Digital Twin Engine.

"These aren't siloed systems," Mr.Dai emphasized. "They operate symbiotically – the Data Engine feeds high-quality corpus to AI LLM Engine, while receiving intelligent data management agents."

The Agent Revolution: ZTE's Co-Sight Framework Leads Industry

From OpenAI's reasoning-focused O1 Model to DeepSeek's multimodel R1 and robotic-integrated Manus and ZTE's GAIA-top-ranked Co-Sight Framework, 2024-2025 marked AI agents' quantum leap.

ZTE's Co-Sight Framework architecture strategically separates Reasoners (analysis) from Actors (execution) to prevent the agent from out-of-focus, with three core innovations driving success: knowledge Graph-optimized data creation, advanced Graph Search for RAG, and standardized LPI Protocols (MCP/A2A) for cross-vendor interoperability.

ZTE had made the Co-Sight Framework open source in April 2025, and with the atomic capability approach, it enables operators to mix-and-match components via OpenAPI or intent-driven interface.

China Telecom Collaboration: Proven Results

As China Telecom's strategic partner, ZTE has collaboratively achieved several significant milestones. These include a 20% reduction in Mean Time to Repair (MTTR), which enables faster troubleshooting. Additionally, ZTE has contributed to a 14.1% improvement in energy efficiency through optimized operations. Furthermore, seamless integration with China Telecom's Kunlun Cloud Network Ecosystem has been realized.

ZTE reaffirmed its commitment to global partnerships, inviting operators to co-create the autonomous network future.