Nokia has named AI-native architecture as the consensus direction for sixth-generation mobile networks, formalising a position that has surfaced in pieces across the industry for the past two years.

The Finnish equipment vendor's declaration argues that AI must be embedded into the radio stack rather than added afterwards, and that core and orchestration layers should start with machine learning from day one.

"AI-native 6G is now the industry direction," Nokia stated.

What does "AI-native" actually mean?

The phrase has been circulating across academic and industry research for several years. Nokia's framing covers three layers: the air interface, where neural receivers and learned channel estimators replace hand-coded signal processing; the radio access network, where AI agents allocate resources and manage handovers; and the operations layer, where model-driven automation replaces today's rule-based network management.

This marks a meaningful shift from 5G, where AI capabilities were bolted on after the standard had been largely finalised. The 5G standard was finished before AI was elevated to a core design principle, and the industry spent the better part of a decade working out how to integrate AI into systems that were never designed for it.

Why is the timing significant?

The declaration arrives as the standards runway for 6G begins to open. The current 3GPP release cycle is closing, and the next study item window for 6G is expected to open in 2025. Vendors that want their architectural assumptions baked into early specifications have to declare them while the study items are still being negotiated.

The wider industry has been converging on similar themes. Several major operators have published research that aligns with AI-native principles, and at least one rival vendor — Ericsson — has publicly framed 6G around cognitive network operations. By claiming a consensus position, Nokia is signalling that this is the direction standards will take unless the rest of the industry works actively to push back.

The financial backdrop matters too. Hyperscalers including Microsoft, Google and Amazon have spent the past two years building partnerships with carriers, packaging the kind of GPU infrastructure that AI-native radio networks will need. Nokia's framing makes those investments look prescient rather than speculative.

What changes for operators and rivals?

For carriers, the practical implication is procurement. Networks that arrive commercially later this decade will, under Nokia's framing, require compute infrastructure capable of running inference workloads at the edge of the radio access network. That shifts capital expenditure from base stations toward GPU-class hardware in distributed sites — a pattern the hyperscaler partnerships have been quietly preparing for over the past two years.

For rival equipment makers, the declaration narrows the strategic space. Vendors that have framed 6G around peak throughput, millimetre-wave economics, or extended reality will have to engage with Nokia's AI framing or risk appearing out of step with the agreed research agenda. Standards-track analysts will watch the next 3GPP meetings for whether the AI-native terminology enters the official study item text.

Smaller vendors will feel the pressure most. Companies that lack an end-to-end portfolio across RAN, core and orchestration will struggle to present a credible AI-native story on their own, raising the prospect of consolidation around the three large RAN vendors that can.

For operators, the shift also redefines the vendor scorecard. Procurement teams have spent the past decade evaluating equipment on peak data rates, spectral efficiency and latency. Under the new framing, the differentiating metrics move toward model training capability, energy use per AI workload, and how cleanly the AI layer interfaces with operations support systems.

What comes next?

The next test for Nokia's AI-native 6G framing is the ITU-R IMT-2030 process, which is due to define a high-level vision recommendation before passing detailed work to 3GPP. Nokia's claim of industry consensus will be validated or weakened depending on how many operators and vendors sign on to its position, and on whether AI-native becomes a normative requirement in the early specifications or remains a research thread.

The Finnish vendor has placed its bet. The rest of the industry now has roughly two to three years to decide whether to fold, fight, or find a third way.