6G news and analysis for the week ending Friday, 9th January, 2026.
Happy New Year everyone and welcome back to 6G Futures. The first full week of 2026 opens with an interesting message: AI is no longer an adjunct to mobile networks but is being designed into their foundations. Three of this week’s most significant stories converge on a single theme — the shift from human-configured network procedures to AI-driven autonomy.
Huawei’s research team has published a detailed blueprint for an “Agentic-AI Core” (A-Core) in the journal Engineering, describing a 6G core network that generates, updates and executes its own control procedures without human intervention. The architecture, built around a large-scale network AI model called NetGPT, treats every service request as a “mission” planned and run by specialised AI agents. This is the most developed version yet of the Agentic-AI concept that Huawei previewed in its 3GPP 6G Workshop submission last year.
Meanwhile, South Korea’s ETRI announced the completion of an AI-based wireless access technology platform (AI-RAN) for 6G. The institute’s Neural Receiver technology demonstrated an 18% improvement in data recovery accuracy in millimetre-wave environments and is now feeding into 3GPP standardisation contributions. ETRI has filed 119 patents and contributed 68 technical articles to 3GPP in this area alone.
Beyond AI, e& in the UAE published a strategic white paper on Integrated Sensing and Communication (ISAC), positioning the technology as a route to “spatial intelligence” where base stations double as environmental sensors. And Airbus is advancing its SpaceRAN demonstrator, testing software-defined 5G satellites that process and regenerate signals on board — a capability essential for the seamless terrestrial-satellite integration that 6G demands. A patent from AT&T covering 6G ubiquitous access was named the top US patent of 2025, underscoring the strategic importance of multi-network convergence.
It is reassuring that we start the year with news that reinforces the direction already mapped out by 3GPP workshop contributors: 6G is likely to be AI-native, sensing-capable and seamlessly connected across land, air and space.
Key announcements this week
ETRI completes AI-RAN platform for 6G
South Korea’s Electronics and Telecommunications Research Institute (ETRI) has announced the completion of its AI-based Radio Access Network (AI-RAN) technology, described as a core foundational technology for the 6G era. The platform applies AI across wireless transmission, network control and edge computing to target up to 10 times higher transmission efficiency than 5G. A standout achievement is the Neural Receiver, an AI-based receiver that directly restores wireless signals and detects errors in real time. Yongsoon Baek, Senior Vice President at ETRI, stated:
“AI-based wireless access technology is the first step in directly implementing the core functions of communication networks through AI.”
What’s new: ETRI’s Neural Receiver demonstrated an approximately 18% improvement in data recovery accuracy, 15% improvement in channel prediction accuracy, and a 30% reduction in data loss rate in millimetre-wave environments compared to conventional methods. The institute has filed 119 domestic and international patents and contributed 68 technical articles to 3GPP, 12 of which were adopted. ETRI plans to develop this into a “Self-Evolving RAN” where the network independently learns and maintains optimal performance.
Why this matters: This represents a substantial step from AI-RAN concept to working platform. While we have seen individual AI-RAN demonstrations from commercial vendors such as KT and Nokia on live 5G networks, ETRI’s work is broader in scope — covering the full chain from beamforming and power control through inter-base-station coordination to edge traffic prediction. The combination of patent filings and 3GPP contributions signals that South Korea is systematically converting research into standards influence. ETRI is also a member of the AI-RAN Alliance, whose Korean membership now includes Samsung, KT, SK Telecom, LG Uplus and the TTA standardisation body.
Implications for 6G standards: ETRI is leading 3GPP standardisation work on “AI/ML-based wireless interface” and “AI-based mobility management.” With 12 contributions already adopted by 3GPP, the institute’s Neural Receiver and Self-Evolving RAN concepts are well-positioned to inform the AI-native RAN specifications expected in Release 20 and 21.
View source material to read more
AT&T 6G patent tops 2025 US patent rankings
AT&T’s patent for a “Framework for a 6G Ubiquitous Access Network” (US Patent No. 12389294) was named the number one patent granted by the US Patent and Trademark Office in 2025, according to IP Watchdog. The patent claims a method of identifying applications on user equipment that can be executed via either terrestrial radio network or satellite network, and reassigning execution to facilitate increased coverage and bandwidth. The invention integrates Wi-Fi, 5G, LTE and satellite networks.
What’s new: The patent describes a framework for dynamic application-level switching between terrestrial and satellite networks to maximise ubiquitous connectivity. It was issued in August 2025 and reflects AT&T’sinvestment in the multi-network convergence that is central to 6G architectural planning.
Why this matters: The recognition of a 6G-focused patent as the top US patent of 2025 signals that intellectual property positioning for 6G is intensifying. AT&T’s approach aligns with the broader industry consensus, as expressed in its own 3GPP 6G Workshop submission, which advocated for a “generation-free” evolution and Sensing-as-a-Service capabilities. The patent’s emphasis on seamless terrestrial-satellite switching is directly relevant to the non-terrestrial network (NTN) integration that virtually every major 3GPP contributor has identified as essential for 6G.
Implications for 6G standards: While patents do not directly determine standards outcomes, they indicate where companies are investing and what they expect to be standardised. AT&T’s patent portfolio in multi-network convergence strengthens its negotiating position in 3GPP, where the interworking between terrestrial and NTN access is a key Release 21 study area.
View source material to read more
e& publishes ISAC white paper for 6G
e& UAE has released a strategic white paper titled “Integrated Sensing and Communication in 6G: e&’s Perspective on Strategic Applications, Monetization Potential, and Global Roadmap.” The paper outlines how 6G infrastructure will evolve from pure connectivity to real-time sensing of the physical world. Marwan Bin Shakar, Acting Chief Technology and Information Officer at e& UAE, said:
“We are moving from a world of simple connectivity to an era of ‘spatial intelligence,’ where the network itself becomes a sensor.”
What’s new: The white paper details monetisation pathways including Sensing-as-a-Service, packaged vertical solutions and SLA-based “safety slices” for reliability-critical services. It ranks market opportunities across manufacturing, transportation, public safety and immersive media. e& UAE references its ISAC drone detection pilot during GITEX 2025 and terahertz spectrum trials as early milestones.
Why this matters: ISAC is one of the defining new capabilities for 6G, featured in virtually every major 3GPP 6G Workshop submission and embedded in the ITU-R IMT-2030 framework. While the technology concept is well established, detailed operator-led analysis of monetisation pathways remains scarce. e& UAE’s white paper adds a Middle Eastern operator perspective to a debate that has been dominated by European and Asian incumbents. The focus on drone detection and autonomous vehicle collision avoidance reflects high-value use cases that could justify early ISAC deployment. The paper follows du’s broader 6G strategy announcement covered in September, suggesting that UAE operators are coordinating their positioning.
Implications for 6G standards: The use cases outlined — including industrial automation, airspace management and privacy-preserving healthcare monitoring — map directly onto the ISAC study items already active in 3GPP. Operator white papers that quantify monetisation potential can help build the business case that influences which features are prioritised in early 6G releases.
View source material to read more
Airbus advances 5G satellite demonstrator for NTN
Airbus UpNext has detailed its SpaceRAN demonstrator programme, which will test a software-defined 5G non-terrestrial network satellite capable of processing and regenerating signals on board. The programme builds on a successful 5G NTN trial completed in February 2025 using Eutelsat’s OneWeb LEO constellation. A ground demonstration will simulate a two-satellite LEO constellation, with an in-orbit demonstration scheduled to launch in 2027 and testing planned for 2028.
What’s new: The SpaceRAN demonstrator goes beyond simple signal relay by implementing on-board signal processing and regeneration. This reduces latency, increases data throughput and enables inter-satellite links for direct traffic routing without ground station intermediation. The consortium includes Aalyria, AccelerComm, CesiumAstro, Deutsche Telekom, Eutelsat, ITRI, Keysight Technologies and others, with the stated goal of creating an open, non-proprietary 5G NTN ecosystem.
Why this matters: The distinction between “bent-pipe” satellites that simply relay signals and regenerative satellites that process them on board is fundamental to NTN performance. On-board processing enables the beam and satellite handovers that are essential for continuous connectivity, particularly for in-flight aviation. Airbus’sown 3GPP 6G Workshop submission advocated for “unification” of terrestrial and non-terrestrial RAT as a viable 6G objective, and specifically called for GNSS-independent operation. The SpaceRAN demonstrator is the practical vehicle for testing these concepts. The emphasis on an open, standardised approach positions Airbus against proprietary satellite connectivity solutions.
Implications for 6G standards: While SpaceRAN targets 5G NTN, its architecture and findings will directly inform the 6G NTN specifications in 3GPP Release 21. The programme’s focus on software-defined satellite payloads, inter-satellite links and standardised open interfaces aligns with the direction mapped by the joint Thales-led 30-company NTN submission to the 3GPP 6G Workshop.
View source material to read more
Selected research papers published this week
Huawei blueprints self-programming 6G core network
A research team from Huawei’s advanced wireless labs in Canada and China has published a detailed blueprint in the journal Engineering for a 6G core network that generates, updates and executes its own control procedures autonomously. The “Agentic-AI Core” (A-Core) architecture centres on NetGPT, a large-scale network AI model fine-tuned on telecommunications knowledge, which interprets natural-language operator intents and chains verified network capabilities into executable workflows.
Why this matters: This is the most detailed public articulation of the Agentic-AI concept that Huawei introduced in its 3GPP 6G Workshop submission (Huawei-250158). The paper describes two concrete use cases — automated device connection and autonomous-driving AI model deployment — and candidly acknowledges open challenges including LLM hallucinations and latency variability. The architecture is presented as a candidate direction for the ITU’s IMT-2030 framework, where “AI and communication” is already an identified usage scenario. The paper’s open-access publication signals Huawei’s intent to build ecosystem support around this approach.
View source material to read more
Agentic AI framework for 6G RAN slicing
Researchers from the University of Ottawa have proposed an Agentic AI framework for 6G RAN slicing that combines a Hierarchical Decision Mamba (HDM) controller with a Large Language Model. The system interprets operator intents expressed in natural language and translates them into coordinated actions across inter-slice, intra-slice and self-healing agents. The paper reports consistent improvements in throughput, cell-edge performance and latency compared to transformer-based and reward-driven baselines.
Why this matters: This paper addresses a gap in current RAN slicing approaches, which typically rely on static mappings or intent-driven reinforcement learning without natural language understanding. The use of Mamba-based architectures (an alternative to transformers with lower computational overhead) is notable for resource-constrained RAN environments. The alignment with the broader Agentic AI trend — where networks interpret human-level intents rather than requiring explicit configuration — echoes the architectural direction proposed by multiple 3GPP 6G Workshop contributors.