6G news and analysis for the week ending Friday, 27th March, 2026.
This week we get the official 3GPP report from its recent TSG#111 plenary meetings in Fukuoka, which have produced the first formal approvals that move 6G from study towards more structured standardisation activity. TSG SA approved the first SA-led 6G study on use cases and service requirements, while TSG RAN’s parallel study on 6G scenarios and requirements has reached 90% completion, and TSG CT approved its first four 6G protocol studies. A cross-TSG session confirmed strong alignment on the Release 21 timeline, clearing the way for a formal announcement at TSG#112 in June. This newsletter has tracked every plenary since TSG#109 in Beijing; and over the weekend we will publish a deeper dive into what actually happened in Fukuoka (as we fight our way around IP blocks...).
The US–China positioning continues at pace. Booz Allen Hamilton made a strategic investment in O-RAN Development Company (ODC), an AI-native Open RAN startup already embedded in the NVIDIA-led AI-WIN initiative. The investment targets defence and allied applications, reinforcing the pivot from traditional Open RAN toward AI-RAN that this newsletter flagged following the NTIA’s Innovation Fund redirection earlier this month. Meanwhile, Chinese state media reports from the National People’s Congress frame 2026 as the critical year for integrating over 300 key technologies into a functional 6G prototype, though the report also acknowledges several shortcomings.
On the infrastructure side, Ericsson had a busy week with two distinct announcements: a multi-year core network modernisation deal with SoftBank in Japan, accelerating the transition to 5G Standalone, and a new research collaboration with Germany’s Forschungszentrum Jülich to explore neuromorphic computing and exascale AI for future networks.
3GPP approves first 6G studies at Fukuoka plenaries
The 3GPP Technical Specification Groups held their 111th plenary meetings in Fukuoka, Japan, from 9 to 13 March 2026, hosted jointly by Japanese organisational partners TTC and ARIB. While 5G Advanced specification work remains the primary focus, the meeting delivered milestone approvals across all three TSGs that mark the formal beginning of structured 6G standardisation activity.
TSG SA approved the first SA-led study on 6G use cases and service requirements (TR 22.870). It is reportedly 590-pages... Come on, who’s going to read through all of that? I mean, my copy of Charles Dickens’ Bleak House is that long! I hope Bleak House isn’t a metaphor here...
Meanwhile TSG RAN’s study on 6G scenarios and requirements (TR 38.914) reached 90% completion. TSG CT approved its first four 6G protocol studies, covering control plane protocols, user plane protocols, resilience and reliability, and NAS protocol, all targeted for completion by September 2027.
A cross-TSG session on the Release 21 timeline confirmed a high degree of alignment between the groups, clearing the way for a formal timeline announcement at TSG#112 in June.
What’s new: SA’s use case report captures the full scope of what 6G is expected to support (and so it should at 590 pages): immersive communication, hyper-reliable low-latency services, massive IoT, ubiquitous connectivity, AI/ML applications and integrated sensing and communications (ISAC). The RAN study is now guiding technical work in RAN working groups on radio service requirements and 5G-to-6G migration, while also shaping 3GPP’s input to the ITU-R’s IMT-2030 technical performance requirements. CT’s four protocol studies represent the first substantive 6G activity in the core network and terminals domain, after initial SIDs slipped from the planned December 2025 target.
Why this matters: Fukuoka was the meeting where the 6G standards process transitioned from scoping to structured study across all domains. The alignment on Release 21 timeline is significant: Previous plenaries saw deliberate caution about committing to dates. The emphasis on CAPEX and OPEX reduction within the SA study also reflects market pressure from operators who need 6G to deliver efficiency gains, not just new capabilities.
Implications for 6G standards: The June TSG#112 meeting is now the decisive milestone. A formal Release 21 timeline announcement will lock in the schedule against which all normative 6G work will be measured. The RAN study’s near-completion means that 6G radio requirements will be available to guide working group activity from mid-2026. CT’s September 2027 completion target for its protocol studies establishes the pacing for core network standardisation.
View source material to read more
Booz Allen invests in AI-native Open RAN for defence and 6G
Booz Allen Hamilton announced that its venture capital arm, Booz Allen Ventures, has made a strategic investment in O-RAN Development Company (ODC), a US-based startup specialising in AI-native Open RAN solutions for 5G and 6G. The investment is part of ODC’s Series A funding round. Chris Christou, senior vice president and 5G/NextG lead at Booz Allen, said:
“We’re advancing AI-native wireless networks that strengthen resilience and bring AI to the edge.”
ODC Chairman Dr. Shaygan Kheradpir described the vision in more expansive terms, claiming the platform can “transform the network from a communication pipe into a distributed compute grid.”
What’s new: The investment connects two threads already visible in US 6G policy. Booz Allen and ODC are both participants in the AI Wireless Innovation Network (AI-WIN), the all-American AI-RAN stack launched last October alongside NVIDIA, SRS and others. ODC was also named in NVIDIA’s MWC 2026 coalition announcement committing to build 6G on open, AI-native platforms. Booz Allen Ventures has a $300 million fund focused on defence tech, AI and deep tech. Initial efforts will target defence applications, with expansion planned into allied international markets, enterprise environments and mobile network operators.
Why this matters: This investment arrives in the context of the NTIA’s recent pivot of the Innovation Fund away from Open RAN toward AI-RAN, covered in last week’s newsletter. The US federal ecosystem is consolidating around a specific vision: software-defined, AI-native wireless networks built on open architectures, with clear defence and national security applications. Booz Allen’s involvement brings government integration expertise that pure-play telecoms startups lack. However, the rhetoric warrants scrutiny. Kheradpir’s claim about transforming cell sites into “high-performance compute hubs capable of orchestrating everything from autonomous agentic AI and real-time generative inference to the physical AI applications” is aspirational and should be read as a vision statement rather than a near-term capability description.
Implications for 6G standards: The investment reinforces the growing US alignment around AI-RAN as the preferred architectural path to 6G. While 3GPP specifications remain implementation-agnostic, the concentration of US defence and venture capital around AI-native Open RAN could influence which technologies US companies champion in the standards process. The connection to NVIDIA’s broader coalition and the OCUDU open-source RAN initiative suggests a coordinated ecosystem strategy.
View source material to read more
China frames 2026 as decisive for 6G prototyping
Chinese state media reports from the National People’s Congress (NPC) highlight 6G as a key future industry alongside quantum technology, embodied AI and brain-computer interfaces. The technology is explicitly named in the 2026 government work report. Miao Wei, NPC deputy and senior vice president at ZTE, said:
“6G represents far more than faster speeds; it’s a leap from simple connection to ubiquitous sensing and intelligent services.”
Professor Zhang Ping from Beijing University of Posts and Telecommunications described the 2025–2026 period as crucial for “integrating over 300 key technologies into a functional, efficient and low-carbon 6G prototype system.”
What’s new: While the broad strokes of China’s 6G programme are well established (we covered the Phase Two trial launch and 300-technology reserve in January) the NPC discussion adds new detail. The government work report’s explicit inclusion of 6G elevates it from a ministry-level programme to a national economic priority. Zhang Ping’s framing of the current phase as moving from “drawing blueprints and gathering materials” to “building the framework and creating the prototype” provides a clearer timeline picture. The report also acknowledges specific weaknesses: critical vulnerabilities in core components and foundational software, and a shortage of interdisciplinary leaders capable of managing the convergence of communications, AI, aerospace and materials science.
Why this matters: The candid acknowledgement of shortcomings is the most interesting element of this report. China holds approximately 40% of global 6G patent applications, but patents alone do not guarantee leadership in the standards process or commercial deployment. The identified gap in interdisciplinary talent reflects a real challenge: 6G requires expertise spanning radio engineering, AI, satellite systems and materials science, a combination that no country currently produces at scale. The call for “Communications + AI” cross-disciplinary programmes signals awareness that the talent pipeline needs structural reform.
Implications for 6G standards: China’s prototype development timeline aligns with the 3GPP Release 20 study phase. System-level integration results from the 300-technology programme will feed directly into contributions from Huawei, ZTE, China Mobile and other Chinese companies during 2026–2027. The emphasis on integrated sensing, AI-native architecture and satellite convergence mirrors the priorities visible across 3GPP working groups, which is unsurprising given Chinese companies’ active participation in shaping those priorities.
View source material to read more
Ericsson modernises SoftBank’s core network for 5G SA transition
Ericsson has signed a multi-year framework agreement with SoftBank in Japan to deploy next-generation core network solutions, expanding and modernising SoftBank’s core infrastructure and accelerating 5G Standalone (SA) adoption. The agreement covers Ericsson’s full cloud-native dual-mode 5G Core solution running on its Cloud Native Infrastructure Solution (CNIS), subscriber data management, policy control and Cloud IMS.
Hideyuki Tsukuda, Executive Vice President and CTO of SoftBank, said that Ericsson’s technologies are “an extremely important foundation for achieving the ‘fusion of AI and networks’ that we are promoting.”
What’s new: The deal is comprehensive. It includes nationwide migration of mobility management functions (MME and AMF) to cloud-native 5G Core, deployment of a full subscriber data management layer through Ericsson’s UDC solution, modernisation of voice and policy control functions, and migration of IMS to Ericsson Cloud IMS. The agreement explicitly targets 5G Advanced use cases and AI-driven operations across the network.
Why this matters: Japan’s transition to 5G SA has been slower than the country’s reputation as a technology leader might suggest. SoftBank’s own 3GPP 6G Workshop submission emphasised that “5GC evolution plus MRSS is the preferred approach” for migrating to 6G, and that NSA is “not preferred.” This core network modernisation is the practical execution of that stated strategy. The cloud-native architecture and automation capabilities being deployed create the foundation layer that 6G will build upon. The deal also reinforces Ericsson’s position in the Japanese market at a time when the vendor is actively signing 6G research partnerships across the region.
Implications for 6G standards: The 3GPP consensus from the 6G Workshop is that 6G will use standalone architecture from day one, meaning operators still running NSA networks likely face a two-step migration. SoftBank’s commitment to cloud-native 5G Core positions it to test and validate the architectural assumptions that underpin 6G standardisation, particularly around AI-driven network management and service orchestration.
View source material to read more
Ericsson and Jülich explore neuromorphic AI for future networks
Ericsson and Forschungszentrum Jülich (FZJ), one of Europe’s leading research centres, have signed a Memorandum of Understanding to develop advanced AI solutions for 5G evolution and 6G. The collaboration will leverage JUPITER, Europe’s first exascale supercomputer, operated by the Jülich Supercomputing Centre. A key focus is neuromorphic computing, which for us mere mortals means brain-inspired approaches to processing complex network tasks with greater energy efficiency.
Prof. Laurens Kuipers, member of the FZJ Executive Board, said the collaboration “has the potential to make a significant contribution to a more sustainable digital future.” Nicole Dinion, Head of Architecture and Technology at Ericsson’s Cloud Software and Services, said the partnership will “explore architectures that define the next generation of telecommunication.”
What’s new: The collaboration covers three research areas: AI methods benchmarked across Ericsson’s full product portfolio using JUPITER for large-scale model training; energy-efficient computing for AI inference at the radio edge, including exploration of neuromorphic approaches such as memristors for tasks like channel estimation and Massive MIMO; and HPC/cloud architecture research drawing on Modular Supercomputing Architecture concepts from the EuroHPC ecosystem. This is distinct from Ericsson’s recently announced Intel collaboration, which focused on silicon and compute architecture, and the VICTOR6G project, which targets industrial 6G applications.
Why this matters: Energy efficiency is becoming a central design constraint for 6G. Neuromorphic computing, which processes information in patterns inspired by biological neural networks, offers potentially dramatic energy savings for the inference workloads that AI-native networks will generate. The connection to JUPITER provides access to computing resources at a scale that most telecoms research programmes cannot match. The collaboration also strengthens European 6G research infrastructure at a time when the EU is actively investing in digital sovereignty.
Implications for 6G standards: While neuromorphic computing is an implementation technology rather than a standards issue, breakthroughs in energy-efficient AI inference could influence what performance levels are considered feasible for AI-native RAN functions. If neuromorphic approaches prove viable for tasks like real-time channel estimation, this could shape the assumptions underlying 3GPP’s AI/ML framework specifications.
View source material to read more
Selected research papers published this week
X-MIMO field trials validate 7 GHz for 6G
Researchers from Samsung, Yonsei University and others present field trial results for eXtreme MIMO (X-MIMO) with 256 digital ports operating in the 7 GHz band. The prototype base station and UE achieved 8-layer downlink single-user MIMO over 100 MHz bandwidth, delivering more than 3 Gbps in urban outdoor environments. System-level simulations compare 128-port and 256-port configurations and examine the design trade-offs involved.
Why this matters: This paper provides independent research validation of the X-MIMO concept that Samsung and KT demonstrated commercially in February. The field trial data confirms that the large digital aperture of X-MIMO supports high-order spatial multiplexing even with limited dominant angular clusters, a finding with direct implications for 6G antenna configuration requirements. The 7 GHz band continues to accumulate real-world performance evidence ahead of WRC-27.
View source material to read more
AI-native control paradigm proposed for fluid antennas
Ian F. Akyildiz and Tuğçe Bilen present a closed-loop AI-native control architecture for fluid antenna networks in 6G. Rather than treating antenna repositioning as a standalone optimisation task, the paper integrates it with conventional radio resource management through multi-agent reinforcement learning. Simulation results show consistent performance gains at the cell edge and reduced inter-cell interference.
Why this matters: Fluid antenna systems, which dynamically adjust physical antenna positions in response to radio conditions, introduce a new degree of spatial freedom for wireless networks. This paper’s contribution is in framing antenna adaptation as a network-level control problem rather than a device-level optimisation, aligning with the broader AI-native architecture vision for 6G. The multi-agent approach addresses practical multi-cell deployment complexity.
View source material to read more
O-RAN security threats mapped for 6G landscape
Lujia Liang and Lei Zhang provide a comprehensive review of security vulnerabilities and privacy challenges in O-RAN architecture as it extends into the 6G context. The paper categorises fundamental vulnerabilities, examines attack vectors and assesses both current and emerging threats. It also reviews existing security mechanisms and ongoing standardisation activities aimed at strengthening the O-RAN security framework.
Why this matters: O-RAN’s open, disaggregated architecture introduces security surfaces that did not exist in traditional integrated RAN deployments. As 6G expands O-RAN capabilities toward terabit data rates, ultra-reliable low-latency communication and non-terrestrial network integration, the attack surface grows correspondingly. This systematic threat mapping provides a reference for both the O-RAN Alliance’s security work and 3GPP SA3’s ongoing 6G security study (FS_6G_Security).
View source material to read more
Aerial agentic AI framework for low-altitude networks
Li Dong and colleagues propose a hierarchical AI framework for low-altitude wireless networks (LAWNs) composed of UAVs and mobile terminals. The system pairs small language models (SLMs) on UAVs for real-time local decision-making with large language models (LLMs) at the base station for deep reasoning and strategy optimisation. A coordination mechanism enables high-frequency local execution with synchronisation to the base station only when necessary.
Why this matters: Low-altitude wireless networks are emerging as a 6G extension layer, but deploying AI capabilities on resource-constrained UAVs presents significant computational and energy challenges. The hierarchical LLM/SLM approach addresses the practical trade-off between intelligence and edge constraints, offering an architectural model that could apply beyond UAVs to other 6G edge computing scenarios.