6G Core Network Functions: Architecture, Intelligence & the Road to AI-Native Networks

Introduction

Every mobile generation has been described in terms of its radio: faster spectrum, wider bandwidth, new antenna techniques. The more consequential change has usually happened in the core. 6G looks likely to follow that pattern, and to push it further. The discussion around future 6G networks is no longer only about connectivity — it spans computing, storage, artificial intelligence, sensing, positioning, edge infrastructure, non-terrestrial access, security and energy efficiency, coordinated as one system rather than as separate domains. The 5G Core introduced a service-based architecture (SBA) in which network functions expose services to each other over standardized interfaces. That foundation is widely expected to be the starting point for 6G, not the endpoint. What is under study is how far the service-based model can be extended when the network must place AI workloads, orchestrate compute, expose sensing data and reconfigure itself continuously.

What Is a 6G Core Network?

In the terms being explored today, a 6G core network is less a fixed collection of network functions and more a control, data-processing, intelligence, security, orchestration and service-enablement layer. Its expected role is to decide not only how a packet reaches a device, but where computation should run, which model should serve an inference request, what sensing data may be shared, how a slice should be composed from multiple resource types, and how all of this should be optimized against latency, trust and energy constraints. The relationship being described in current research and standards discussion is a convergence: communication plus computing plus storage plus AI/ML plus sensing plus positioning plus edge plus non-terrestrial access plus applications plus security plus automation, coordinated by a common control and exposure layer.

6G Core Network Functions at a Glance

1. Access and Mobility Management

Mobility in a 6G context is expected to be harder than in any previous generation, because the set of things a device may move between is much larger. Future access and mobility management may need to handle movement across terrestrial networks, non-terrestrial networks including satellites and UAV platforms, edge networks, and multiple radio technologies operating in the same environment. In a constellation-based non-terrestrial deployment, the cell itself moves; in dense terrestrial deployments, the device moves through highly heterogeneous coverage. The architectural direction being explored is mobility that is predictive and context-aware rather than purely reactive — anticipating handovers, pre-positioning session state, and considering compute and application anchoring alongside radio continuity.

2. Advanced Session Management

Session management in the 5G Core already goes well beyond bearer setup, but future session management may take on a much wider set of inputs: latency targets, reliability requirements, compute location, AI inference needs, data locality and residency, energy cost, sensing requirements, security posture and slice membership. The direction of travel is from connection setup toward service- and resource-aware session orchestration — where the network decides jointly what path the data takes and where the processing happens.

"Service- and resource-aware session orchestration"

3. Distributed User Plane Functions

The user plane is the clearest example of decentralization. UPF-like processing is expected to move closer to users, devices, vehicles, industrial systems, edge applications, sensors and AI inference resources. Instead of a small number of large aggregation points, the user plane may become a distributed fabric whose elements are instantiated, discovered and composed where they are needed. That has second-order effects: forwarding decisions become entangled with compute placement, and mobility becomes a question of moving processing as well as traffic.

4. AI and Machine Learning Functions

The phrase AI-native 6G describes something narrower than 'the network uses AI'. It describes AI as part of the network control loop itself. Potential AI/ML functions discussed in research and standardization contexts include network-state prediction, traffic prediction, resource optimization, anomaly detection, energy optimization, slice optimization, mobility prediction, AI model deployment and lifecycle management, distributed inference, federated or distributed learning, policy assistance and — at the far end of the spectrum — autonomous network control. Whether and how these become standardized functions remains open.

5. Network Data and Analytics Functions

Data functions are the substrate on which AI-native behaviour depends. The scope being discussed covers network performance, traffic, devices, applications, resources, QoS, energy, security posture, mobility and slice performance. In the 5G Core ecosystem, NWDAF represents the current direction for network data analytics; it is part of an evolving 5G Core, not a finalized 6G function. What changes in a 6G context is the expectation of timeliness: analytics that inform planning are useful, but analytics that close a control loop in near real time are a different architectural requirement, with implications for where data is stored, how it is exposed, and how privacy is enforced.

6. Dynamic Network Slicing

Slicing today is largely a connectivity construct. The emerging direction is a slice that coordinates connectivity, compute, storage, AI resources, edge capacity, security properties, sensing capability, energy budget and transport. A slice, in that reading, becomes a service contract across multiple resource types that can be composed, monitored and re-optimized during its lifetime rather than configured once.

7. Policy Control

Policy control is where intent becomes behaviour. Future policy decisions are expected to be context-aware, AI-assisted, application-aware, slice-aware, energy-aware, security-aware and location-aware — evaluated continuously rather than provisioned statically. The practical challenge is explainability: a policy engine that adapts autonomously must still produce decisions that operators can audit, especially where regulation, safety or contractual SLAs apply.

8. Authentication, Identity and Trust

Identity in a 6G core may need to cover far more than subscribers. The entities requiring identity and trust could include devices, AI agents, network functions, applications, data sets, models, edge resources and virtualized network components. Distributed deployment compounds the problem: when functions are instantiated dynamically across operator, cloud and edge domains, trust cannot be assumed from topology.

"How is trust established when autonomous AI agents and distributed network functions make decisions?"

9. Network Exposure and API Functions

Network exposure is where 6G becomes commercially interesting to industries outside telecom. The direction under discussion evolves NEF-style exposure toward broader programmable network capability. Potentially exposed capabilities include connectivity, QoS, positioning, sensing, network slices, edge computing, AI inference and analytics — each with authorization, privacy and charging implications.

10. Compute and Edge Orchestration

Compute-aware networking is one of the most distinctive shifts. Rather than requesting bandwidth, an application may request a coordinated bundle of resources — for example 1 ms latency plus compute capability plus AI inference plus geographic constraints plus security properties plus reliability. Satisfying such a request requires the core to select edge sites, place workloads, route traffic and reserve capacity as one decision.

11. Integrated Sensing and Communication

Integrated sensing and communication (ISAC) uses the same radio infrastructure for communication and for sensing the physical environment — object detection, localization, imaging, mapping and environmental awareness. The core's potential role is governance and coordination rather than signal processing: managing sensing policies, sensing data flows, application access, privacy constraints, positioning services and edge analytics. Privacy is not a footnote here; a network that can sense its environment raises consent and data-governance questions that architecture must answer explicitly.

12. Positioning and Location Intelligence

Positioning is expected to become an intelligence service rather than a coordinate lookup, drawing on location services, positioning data, mobility context, sensing information, application requirements and privacy policies. Use cases driving this include autonomous vehicles, robotics, industrial automation and immersive systems, where position, timing and reliability are coupled to safety.

13. Energy and Sustainability Functions

Energy is emerging as a first-class control objective. The direction discussed is dynamic optimization of compute placement, network resources, data paths, AI workloads, network slices, edge infrastructure and sleep modes — with the goal of meeting service requirements while minimizing energy consumption, rather than treating efficiency as a deployment-time configuration.

14. Distributed and Self-Organizing Core Functions

The final architectural shift is away from the idea of the core as a centralized collection of network functions. Emerging research describes core functionality that is distributed across sites and domains, discovered dynamically, composed into service chains, orchestrated against intent, optimized continuously and reconfigured as conditions change. Self-organizing core concepts remain research directions and should be read as such — the operational, assurance and regulatory questions they raise are unresolved.

A Conceptual 6G Core Network Architecture

The layering below is a conceptual synthesis of the directions described above. It is offered as a way to reason about where innovation may concentrate — not as a depiction of any official architecture.

6G Core vs 5G Core

"6G is not simply “5G Core with more network functions.”"

Why 6G Core Functions Matter for Patent Strategy

When a new set of technical problems must be solved for the first time, the resulting solutions tend to be captured in patent filings well before they appear in a specification. That is the practical significance of an architecture still under definition: innovation opportunities can emerge across multiple layers simultaneously, and the layers are not equally crowded. The categories below map where filing activity around future 6G core functionality may concentrate. None of this implies that any particular filing is novel, granted, essential or enforceable — those are separate determinations requiring claim-level analysis.

What Companies Should Monitor in 6G Core Patent Landscapes

A landscape built around a moving standard needs a stable frame. An eight-layer framework — core architecture, control functions, user plane, AI/ML, compute, sensing, security and exposure — gives one, because each layer maps to a distinct family of technical problems rather than to terminology that may be renamed as standardization proceeds. Within each layer, a landscape should examine technology white spaces, competitor concentration, emerging assignees, potential freedom-to-operate issues, and the split between architectural innovation and implementation-level innovation.

One methodological point matters more than the rest. Searching only for identical terminology will systematically miss relevant art, because pre-standard filings rarely use the vocabulary the standard eventually adopts. Effective searching examines architecture, functional relationships, signalling procedures, resource allocation mechanisms and technical effects — the invariants that survive renaming. This is the same discipline that underpins credible patent-to-standard mapping work, and it is where technology intelligence and IP intelligence become the same exercise.

The Most Important 6G Core Trend

The properties that follow from that shift — AI-native, distributed, programmable, autonomous, compute-aware, sensing-aware — are the characteristics most likely to define which technical contributions matter as the architecture settles.

Conclusion

The 6G core is still being defined. What is visible today is a direction rather than a specification: a distributed, intelligent and programmable system coordinating connectivity, computing, AI, sensing, security, resources and services as one control problem. For organizations building in this space, that ambiguity is the opportunity. Patent strategy, competitive intelligence, freedom-to-operate analysis, technology landscape analysis and portfolio planning are all more tractable before a standard consolidates than after — provided the analysis is anchored in architecture and technical effect rather than in vocabulary that has not stabilized yet.