Seventeen major mobile operators filed a joint position paper ahead of 3GPP's Madrid meeting this autumn, demanding evolutionary, upgradable paths to 6G rather than rip-and-replace core reinvention. The operators still carry capital costs from 5G standalone deployments that have not yet produced substantial commercial returns.

The paper, written by major Western carriers, signals a clear preference for an evolutionary 6G core anchored to the 5G service-based architecture. It comes as the standards body's SA2 working group meets to consolidate competing designs submitted earlier in the cycle.

What do the three candidate directions look like?

3GPP has narrowed 6G core proposals to three candidate directions, all centered on how to embed AI and rework non-access stratum (NAS) signaling:

• Direction 1 – Separate AI domain: A dedicated AI domain for intent handling, independent of the packet-switched network, with the AMF routing NAS signaling while AI fulfillment occurs via service-based interfaces. Backed by Nokia, T-Mobile USA, Verizon, NVIDIA, Deutsche Telekom, Qualcomm, Ericsson, NEC and Apple, which also fits Direction 2 because it does not specify AI realization. Requires complex UE AI domain client integration.
• Direction 2 – AI functionality in 6G NFs: Integrates AI through dedicated 6G network functions following standardized 3GPP procedures with flexible NAS routing (standalone or combined with core functions like the AMF). Supported by NTT DOCOMO, Samsung, LG Electronics, NEC, IIT Bombay, Vodafone, Apple and China Telecom, partly.
• Direction 3 – AI/agent-handled connectivity: Deeply integrates AI and agents into network procedures with dynamic coordination of capabilities and tool invocation, using a signaling routing function independent of the AMF. Aligned with Huawei, HiSilicon, China Mobile, ZTE, vivo, CATT, ETRI, Ewha Womans University and OPPO, partly. Introduces high procedural disruption.

Direction 1 has gained the most traction among Western operators and vendors. Anchoring the core to the 5G service-based interface preserves 5G core investments while letting intent handling mature alongside existing services.

Why does NAS signaling matter so much?

Non-access stratum is the secure signaling pipeline between device and core network, handling authentication, encryption, mobility management and session setup. Mobile architectures have relied on it since 2G (GSM), embedding it in device baseband silicon and SIM security frameworks.

Reworking NAS requires modem redesign, core security procedure overhaul and device backward-compatibility planning. Any 6G decision that breaks NAS therefore reaches into every handset, baseband supplier and SIM on the planet.

Is a compromise taking shape in Prague?

Delegates floated a partial merger of Directions 1 and 2. Operators would keep a separate AI domain but introduce modular routing to prevent potential chokepoints in the AMF. In practice, this means tweaking NAS routing so it can operate standalone or combine with core functions like the AMF.

The compromise still favors Direction 1's commercial logic: investment protection, multi-vendor interoperability and a smooth migration from 5G SA. It also leaves room for Direction 2's standardized network functions.

What happens at the device edge?

Even if 3GPP converges on Direction 1, architectural disruption moves out to the terminal in the form of the UE AI domain client. Direction 1 standardizes client interfaces to the core by creating an application-layer intent client and a 3GPP AI domain client.

Client placement carries commercial weight:

• Baseband firmware: tightest 3GPP integration, but tied to silicon cycles at Qualcomm, MediaTek and Samsung.
• Operating system layer: faster software updates via Android and iOS, though it requires new APIs to translate application intent to baseband hardware.
• App layer: quickest deployment, but lacks the low-level integration needed for deterministic performance.

If specifications lean toward an operator-controlled framework, device-side AI agents, hyperscalers and third-party over-the-top ecosystems risk lock-out. An OTT AI agent running cloud gaming or enterprise workloads might be forced through an additional operator translation function to dynamically adapt network parameters.

Could the global standard splinter?

In previous generations, 3GPP could resolve deadlocks by leaving contentious variations as optional vendor implementations outside 3GPP specifications. For 6G, architectural divergence is too deep for that workaround.

Direction 3 represents a revolutionary clean-slate design with native agentic routing and independent signaling functions. The other two directions remain evolutionary, but the gap between these cohorts could leave operators managing two divergent cores: a single-vendor revolutionary stack in Chinese domestic markets and a multi-vendor evolutionary architecture elsewhere.

What is the next milestone?

Post-Madrid discussions must finalize a 6G core architecture by November. The deadline is tight for a system solid enough to keep a divided industry unified under a single global standard, and how 3GPP balances operator control with device ecosystem openness remains unresolved.