The operation is already profitable on paper. NVIDIA agreed in October 2025 to invest $1 billion in Nokia at $6.01 per share, a placement that gave it approximately a 2.9% stake in the company. On Friday, April 10, 2026, Nokia’s ADRs closed at $9.46 in the U.S., representing an appreciation of nearly 57.4% from the initial NVIDIA purchase price.
But reducing the deal to a mere capital gain would be superficial. The truly important aspect is that the partnership between Nokia and NVIDIA indicates a structural shift: telecommunications networks are no longer viewed solely as channels for carrying traffic but as a future distributed layer of AI inference. This is the point that begins to change how Nokia is perceived from a technological—and partly financial—perspective.
At the end of October 2025, NVIDIA and Nokia announced a strategic alliance to add AI-RAN commercial products to Nokia’s radio access portfolio. The clear thesis: bring AI workloads to the network edge, where latency truly matters, using existing telecom infrastructure. During the announcement, Nokia went so far as to say they aim to “put an AI data center in everyone’s pocket,” a phrase capturing the industrial ambition of the project well.
The most eye-catching aspect is the use of accelerated hardware directly within network environments. At GTC 2026, NVIDIA explained that its AI-RAN lineup includes NVIDIA ARC-Pro, based on the RTX PRO 4500 Blackwell Server Edition for energy-constrained cell sites, and RTX PRO 6000 Blackwell Server Edition for higher-capacity mobile offices. NVIDIA also confirmed that T-Mobile was the first carrier in the U.S. to pilot this AI-RAN infrastructure with Nokia’s anyRAN software.
An important clarification: market commentary often talks about “towers” in cellular deployment as if all infrastructure will be located solely at the antenna mast. The official reality is somewhat broader. NVIDIA refers to “AI grids” built on a distributed network footprint that includes regional hubs, mobile switching offices, and central offices, not just tower sites. According to the company, there are approximately 100,000 distributed network centers worldwide with enough spare capacity to add over 100 gigawatts of new AI capacity over time.
This significantly changes the framework. It’s no longer just about upgrading mobile networks but about repurposing existing physical infrastructure—ground, power, connectivity, metropolitan locations—into near-user, near-device or factory inference platforms. For use cases such as robotics, industrial automation, artificial vision, or physical AI, the advantage of running inference close to the data source is clear: less latency, reduced back-and-forth to the cloud, and potentially more cost-effective per token in certain scenarios. That is the promise.
Nokia is not only betting on AI-RAN: it also has an optics strategy
The other key angle of the thesis lies in optical networking. Nokia completed its acquisition of Infinera in February 2025, strengthening its position in optical networks and among hyperscalers precisely as demand for AI infrastructure began to push backbone and interconnection capacities.
The latest data from Nokia supports this direction. In Q4 2025, the company reported that Optical Networks grew 17%, and for the full year, it highlighted €2.4 billion in orders from AI & Cloud clients. Additionally, in November 2025, Nokia announced a new strategy aiming to increase its operating profit to €2.7–3.2 billion in 2028, up from a baseline close to €2 billion.
This move is not isolated. Ciena, a major comparable in the optical space, projected revenues of $5.7–6.1 billion for fiscal 2026 in December 2025, and by March 2026, raised that outlook to $5.9–6.3 billion, also reflecting the surge in demand driven by AI and advanced network infrastructure.
What is confirmed and what remains interpretive
There are several solid facts. First, NVIDIA acquired a 2.9% stake in Nokia at $6.01 per share. Second, both companies are pushing AI-RAN forward, and T-Mobile has already piloted it with Nokia’s anyRAN software. Third, NVIDIA sees a huge opportunity in distributed networks as the future layer for AI inference. Fourth, Nokia’s optical division is benefiting from hyperscalers and AI & Cloud spending.
What is not officially confirmed is the more aggressive market interpretation—that “NVIDIA did not buy Nokia for exposure in the stock market but for distribution.” While this could be a reasonable strategic hypothesis, it remains an inference, not a formal statement from NVIDIA. The companies have only said that their alliance aims to accelerate AI-RAN, AI-native 5G/6G, and new telecom architectures supported by NVIDIA’s accelerated computing and Nokia’s network stack.
Why this story matters more than it might seem
For years, the market has primarily viewed Nokia as a cyclical telecom supplier highly exposed to 5G investments. The new narrative emerging now is that Nokia is becoming one of the few Western players integrating RAN, optics, software-defined networking, and operator relationships at a time when AI aims to emerge from data centers and dominate network edges.
This does not mean the shift is guaranteed nor that AI-RAN will materialize quickly. Real deployment will depend on monetization, energy costs, software maturity, operator acceptance, and competition from other suppliers. But it does indicate that telco networks are beginning to be seen less as passive infrastructure and more as distributed computing platforms. In this context, Nokia has successfully positioned itself within a conversation previously dominated almost entirely by hyperscalers and chip vendors.
FAQs
How much has NVIDIA earned from its investment in Nokia so far?
Based on the subscription price of $6.01 in October 2025 and the $9.46 closing on April 10, 2026, the appreciation is approximately 57.4%.
What exactly is AI-RAN?
It’s an architecture that integrates AI workloads into the Radio Access Network, allowing the same infrastructure to support network functions and perform AI inference at the edge. Nokia and NVIDIA are presenting it as a key step toward advanced 5G and native AI-oriented 6G networks.
Is T-Mobile already using this technology?
T-Mobile was the first U.S. carrier to pilot NVIDIA’s AI-RAN infrastructure with Nokia’s anyRAN software, and has participated in over-the-air tests and demonstrations of concurrent AI and RAN processing.
What role does Infinera play in this story?
The acquisition of Infinera strengthened Nokia’s optical business, an area witnessing increased demand from hyperscalers and AI & Cloud clients. Nokia ended 2025 with €2.4 billion in orders in that segment and saw strong growth in Optical Networks during the fourth quarter.
via: X Twitter