RAN Architecture for the AI-Native 6G Era

Why It Matters

An AI‑native, open RAN is essential for operators to deliver ultra‑low‑latency services, control costs, and capture emerging 6G revenue opportunities.

Key Takeaways

• AI-native RAN must integrate GPUs, not bolt‑on, for deterministic performance.
• Operators need software‑defined, open architectures to avoid proprietary lock‑ins.
• Edge inference moves compute closer to users, enabling real‑time services.
• Automation and lifecycle management are critical to control rising operational complexity.
• Dynamic traffic patterns demand autonomous networks for cost‑effective scaling.

Summary

The RAN Summit panel explored how 6G will be built on an AI‑native radio access network, with Guy Daniels hosting Alisa’s COO Sami Ko‑Linan and Wind River’s VP Warren Beck. They argued that the next‑generation RAN cannot treat AI as an afterthought; instead, GPUs and inference engines must be baked into the core architecture to preserve deterministic performance required for carrier‑grade deployments. Key insights included a shift from hardware‑centric designs to intelligence‑driven, software‑defined networks, open‑infra interfaces, and distributed edge compute. Operators must adopt autonomous orchestration that moves workloads and data to the optimal point—whether at the base‑station, local breakout, or cloud—while maintaining isolation and SLA guarantees. The panel highlighted the importance of lifecycle automation to tame the operational complexity introduced by Open RAN disaggregation. Notable remarks underscored the strategic pivot: Warren warned, “AI is not a bolt‑on, it’s part of the architecture,” and Sami emphasized moving away from proprietary black‑boxes toward open, cloud‑native platforms. Real‑world pilots with Wind River’s local breakout solution and partnerships with Google and Nokia illustrate how edge inference can unlock ultra‑low‑latency, monetizable services beyond traditional mobile broadband. The implications are clear: operators who invest now in open, AI‑centric RAN, GPU‑enabled hardware, and automated lifecycle management will secure cost‑effective scalability, faster time‑to‑market for new services, and new revenue streams in the 6G era.

Original Description

Sami Komulainen of Elisa and Warren Bayek of Wind River discuss the concept of 6G shifting from hardware-reliant ‘black boxes’ to autonomous, cloud-based architectures. By leveraging 5G foundations, edge inference, and digital twins, this intelligence-driven approach reduces latency and operational complexity. This panel discussion looks at how innovations aim to bring compute power closer to users, enabling rapid service launches and new monetisation through device-to-device communication.

Featuring:

Sami Komulainen, Chief Operating Officer, Executive Vice President, Technology and Operations, Elisa

Warren Bayek, VP, Intelligent Edge, Wind River

Recorded: March 2026

#telecomtv #ran #radioaccessnetwork #elisa #windriver #6g #ainative