Nokia AI-RAN Trials Expand to Eight Operators Globally as U.S. Carriers Diverge

Nokia AI-RAN Trials Expand to Eight Operators Globally as U.S. Carriers Diverge

Nokia has confirmed that eight international operators are now running active AI-RAN proofs of concept on Nvidia's Aerial compute platform, while America's three major carriers — AT&T, T-Mobile, and Verizon — are each pursuing their own separate AI radio network strategies. The divergence signals that the path to AI-native wireless infrastructure will be uneven, with international and Tier 2 operators potentially leading the deployment curve.

What is AI-RAN?

AI-RAN (AI Radio Access Network) is an architectural approach where machine-learning models replace or augment traditional signal processing functions in mobile base stations. Instead of fixed algorithms handling spectrum use and interference management, AI-RAN systems adapt in real time — improving throughput, reducing energy consumption, and enabling dynamic resource allocation across thousands of connected devices simultaneously. Nokia has partnered with Nvidia to bring this capability to its AirScale platform using Nvidia's Aerial software stack, which runs radio processing workloads on GPU hardware rather than purpose-built ASICs (application-specific integrated circuits — chips designed for a single task).

The Eight Operators

According to Nokia's September update, the operators running active AI-RAN proofs of concept as of mid-September 2026 are:

• A1 Group (Austria and Eastern Europe)
• Chunghwa Telecom (Taiwan)
• du (United Arab Emirates)
• e& (Abu Dhabi)
• Mobily (Saudi Arabia)
• stc (Saudi Arabia)
• TPG Telecom (Australia)
• Zain Saudi (Saudi Arabia)

The concentration of Gulf-region operators — four of the eight — reflects the rapid 5G buildout underway in Saudi Arabia and the UAE, where greenfield deployments face fewer legacy integration constraints than mature Western markets. These carriers are not ripping out and replacing existing infrastructure; they are testing AI-RAN in discrete network segments alongside conventional radio deployments.

Why U.S. Carriers Are Moving Differently

AT&T, T-Mobile, and Verizon are each taking separate paths to AI integration in their radio networks. AT&T has oriented toward open RAN architectures that allow software from different vendors to run on shared hardware. T-Mobile has focused on optimizing its existing multi-vendor supply chain. Verizon has pursued targeted proprietary partnerships rather than broad platform bets. None of the three are currently participating in Nokia's Nvidia-based AI-RAN proof-of-concept program.

The divergence is not surprising. Large U.S. incumbents carry tens of billions of dollars in legacy radio equipment and serve hundreds of millions of customers on networks where a proof-of-concept failure has real operational consequences. Smaller international carriers with newer infrastructure can afford to experiment in ways U.S. Tier 1 operators cannot.

What This Means

For Nokia: The company is building an AI-RAN reference customer base outside the United States that it can use to pressure American carriers with proven performance data. Each successful POC becomes a sales tool for the next conversation. Ericsson and Samsung are running parallel AI-RAN programs, so Nokia's urgency in expanding its operator list is competitive as much as technical.

For Nvidia: Every AI-RAN proof of concept is a GPU foothold inside telecom infrastructure — a market that has historically run on purpose-built chip hardware. Eight operators simultaneously running Aerial workloads validates the platform's commercial applicability beyond data centers and signals a potential new revenue stream for Nvidia's networking business.

For enterprise buyers: AI-native networks promise lower-latency, higher-reliability connectivity for industrial applications — factory automation, logistics robotics, connected vehicles, real-time remote operations. If international operators prove measurable ROI at scale, enterprise customers in the United States will eventually push their carriers to match those capabilities or lose contracts to providers that can.

What to Watch

Nokia will need at least one of its eight POC operators to commit to a full commercial AI-RAN deployment before year-end to demonstrate that the technology crosses from lab to live network. On the U.S. side, the carriers' divergent strategies will eventually reach a consolidation point — most likely forced by an enterprise customer demanding AI-RAN service-level agreements that existing architectures cannot meet.