Observed Signal · Oct 16, 2026 · Technical Release · Source: PR Newswire: Advertising & Marketing · Impact: 4/5 · Sentiment: Positive
Building AI-Native Core for Next-Gen Agentic Services
Huawei outlines a vision for an AI-native core network architecture designed to support next-generation AI-agent services. The company predicts that by 2035, there could be 900 billion AI agents worldwide, necessitating a shift from bit-oriented to token-oriented traffic handling. Huawei identifies three typical service quality scenarios: AoNR for low-latency task execution, Wide-Area Computing Network (WCN) for improved computing efficiency, and Agent Computing Network (ACN) for secure collaboration among AI agents. The new core network should feature intent-based interfaces, a dedicated Data & AI Plane (DAP) alongside traditional control and user planes, and deep integration of AI decision-making into network procedures. Huawei outlines two evolution phases: first, adding AI capabilities as overlays, and second, creating a fully AI-native core network capable of enhanced connectivity and beyond connectivity services, requiring ecosystem collaboration across devices, chips, and networks.
Huawei's vision for an AI-native core network is a major platform announcement that outlines the future infrastructure for AI-agent services, which will be crucial for the AdTech ecosystem as it moves towards agentic advertising and autonomous media buying.
Track Huawei Signals & Market Shifts in Real-Time
Polaris7 autonomous intelligence agents track regulatory filings, primary sources, executive changes, and deal flow 24/7. Create your free Explorer workspace to monitor these entities.
Key Takeaways & Evidence Grounding
- Huawei predicts 900 billion AI agents globally by 2035.
- Huawei introduces three service quality scenarios: AoNR, WCN, ACN.
- New core network architecture includes intent-based interfaces and a Data & AI Plane (DAP).
- Huawei outlines two evolution phases towards an AI-native core network.
Connected Companies & Entities
1 Entity mapped“Huawei believes that an AI-native core network will be crucial for this transformation.”
Ontology Mapping & Concepts
Related Market Signals & Shifts
Recent verified developments and strategic activity across this market segment.
Huawei President Urges Network Rebuild for AI Era
At the Network X 2026 conference in Vienna, Bob Chen, President of Huawei's Optical Business Product Line, delivered a keynote at the WBBA's Broadband Development Congress, advocating for a network transformation to meet AI era demands. He proposed three 'R's: Rebuilding network architecture around AI Data Centers (AIDCs) instead of administrative regions, with latency guarantees of 1-5-20 ms; Rethinking business models to offer 'bandwidth + latency + reliability' and integrated network-compute-model services, expanding home broadband from connectivity to whole-home intelligence; and Redefining network operations through full-lifecycle AI, including intelligent Wi-Fi tuning and energy-saving mechanisms. Chen highlighted that AI large models now have nearly 10 trillion parameters and token consumption has increased almost 50x in the past year, necessitating these shifts. AI-driven tuning can boost speeds by ~20% and reduce device energy by over 20%.
Huawei Predicts 900 Billion AI Agents by 2035
Huawei's Intelligent World 2035 report forecasts 900 billion AI agents by 2035, one hundred per human, with agents moving 90% of AI token traffic and driving a 100-fold rise in data volume. Reuters corroborated the agent traffic prediction. Meanwhile, hyperscalers have committed $725 billion in capex for 2026, up 77% year-over-year. The article interprets this as a bet on machine-to-machine commerce, where agents act as customers for other agents, rather than tools for humans. It highlights the trend of AI agents as economic actors, potentially reshaping the internet and advertising ecosystems.
Agentic AI Transforms Telecom Network Management
This developer guide explains how 'agentic AI'—autonomous agents that perceive, reason, act and learn—are being applied to telecommunications network operations. It describes a common perception→reasoning→action→learn architecture and concrete observation and action-layer technologies (gNMI/gRPC, SNMP, Netflow, Kafka/Pulsar, InfluxDB/VictoriaMetrics, Neo4j, SDN APIs, Ansible/Terraform). Key production use cases include autonomous fault remediation (with Telefónica citing MTTR reductions >50% in some fault categories), predictive capacity management, RAN self-optimization (O‑RAN xApp/rApp) and network slice orchestration. The article stresses data‑engineering effort (40–60% of initial project work), safety controls (blast-radius limits, reversibility, dry-run, escalation), phased deployment (monitor-only → supervised automation → full autonomy) and a practical multi‑quarter rollout plan. It notes tooling (vector DBs for RAG, RL, GNNs and LLM-based reasoning) and flags multi-agent coordination and O‑RAN ecosystem maturity as the next challenges.
Receive all future signals for Huawei & competitors in your briefing
Never miss material M&A deals, leadership changes, account wins, or product rollouts. Polaris7 monitors 20,000+ primary sources 24/7, delivering verified telemetry directly into your executive briefing.