Nokia bets on Nvidia AI-RAN and Microsoft automation to pivot toward 6G

Key points

• Nokia is integrating its AI-native anyRAN software with Nvidia’s Aerial RAN Computer, claiming over 20% improvements in spectral efficiency through AI-driven radio innovations.
• The company announced a collaboration with Microsoft to combine its Data Suite with Microsoft Fabric, enabling agentic network automation for tasks like autonomous Voice over New Radio assurance and geo-experience analysis.
• Nokia is expanding AI-RAN trials across North America, Europe, Asia-Pacific, and the Middle East, with operators including A1 Group, Chunghwa Telecom, and stc participating in proof-of-concept deployments.
• Nokia’s strategy includes a new 'dApp' interface (E3) for real-time network control, which Ericsson and Samsung have declined to support, preferring their own proprietary baseband solutions.
• Financially, Nokia reported a 105% year-over-year increase in sales to AI and cloud customers in Q2 2026, driven by demand for optical networking and IP infrastructure for AI data centers.

Background

Nokia has been losing market share in the US and globally, prompting a strategic shift under CEO Justin Hotard. The company acquired Infinera in 2025 to bolster its photonics and optical networking capabilities, which are critical for AI data center interconnects. This move follows a broader industry trend where AI is moving from an application layer to a native component of network infrastructure, aiming to reduce reliance on hardware upgrades for future 6G capabilities.

How outlets are covering it

Light Reading highlights the risk in Nokia’s bet on Nvidia GPUs, noting that while 19 entities have signed up for initial AI-RAN deployments, many are startups rather than major telcos. It also points out that Ericsson and Samsung are resistant to xApps and dApps, viewing them as threats to their proprietary baseband technology. Konsulteer emphasizes the positive financial impact, citing a 105% surge in AI/cloud sales and the strategic value of the Microsoft partnership for data governance. Pluang focuses on the broader pivot to AI infrastructure, noting that while RAN growth is slow, optical and IP network segments are expanding rapidly due to AI data center demand, despite recent revenue declines and cash flow challenges.

Why it matters

Nokia’s strategy could redefine the telecom industry by making AI a core component of network infrastructure rather than an add-on. If successful, it may force competitors like Ericsson and Samsung to adopt similar open RAN standards, potentially disrupting the current vendor-locked market. However, the success of this pivot depends on telcos adopting GPU-based RAN, which remains a significant gamble given the current dominance of traditional silicon in the industry.

What to watch

Nokia plans to further expand AI-RAN trials and integrate dApps for real-time network control by 2027-2028. The company will continue to push for spectral efficiency improvements and autonomous network operations, while also expanding its AI infrastructure partnerships, including the Zankore initiative in Asia-Pacific. The next few years will determine whether telcos will embrace GPU-based RAN or stick with traditional vendor-locked solutions.