Nokia Launches NVIDIA-Powered AI-RAN Platform, Betting on AI to Revive Its Mobile Network Business
Nokia has officially launched its new AI-RAN platform, combining its anyRAN software with NVIDIA’s Aerial technology to help telecommunications operators increase network capacity without immediately purchasing additional spectrum.
Announced on July 15, Nokia described the platform as the industry’s first GPU-accelerated AI-RAN solution. The company claims that artificial intelligence can significantly improve how mobile networks manage radio frequencies, traffic, and computing resources.
According to Nokia, early testing has already demonstrated spectral efficiency improvements of more than 20 percent. The company is targeting gains of 50 percent by 2027 and more than 100 percent by 2028.
If Nokia achieves those targets, telecom operators could potentially double the capacity of their existing spectrum.
However, the larger performance figures remain future targets rather than verified commercial results. Nokia plans to begin operator pilots toward the end of 2026, with full commercial availability expected in 2027.
AI-RAN Through Software, Not a Complete Hardware Replacement
One of Nokia’s main selling points is that operators may deploy the technology through a software subscription instead of replacing their entire network infrastructure.
The company is offering three deployment options:
- A GPU-powered plug-in card for existing Nokia AirScale sites
- A standalone AI-RAN network node
- A cloud-based server deployment delivered through technology partners
This approach could allow telecom providers to modernize their networks gradually while extending the usefulness of existing radio equipment.
Unlike traditional radio access networks, AI-RAN combines mobile network processing and AI workloads on shared computing infrastructure. The goal is to use artificial intelligence to optimize network performance in real time while also allowing operators to run AI applications closer to users.
Nokia’s Attempt to Rebuild Its Mobile Network Business
The launch is more than a technology announcement. It is also part of Nokia’s effort to revive its struggling mobile networks division.
Since becoming Nokia’s chief executive in 2025, Justin Hotard has identified the company’s mobile radio business as one of its most difficult challenges.
During Nokia’s capital markets presentation in November, Hotard told investors that the mobile business had failed to deliver acceptable returns. The company later reorganized the unit under a new Mobile Infrastructure segment and introduced additional cost-reduction measures.
Nokia’s partnership with NVIDIA now sits at the center of that turnaround strategy.
In October 2025, NVIDIA reportedly invested approximately $1 billion in Nokia in exchange for a stake of around three percent. By using NVIDIA’s GPUs and CUDA software, Nokia can reduce some of the cost and complexity involved in designing its own specialized computing chips.
The strategy also supports Nokia’s transition from a hardware-heavy business model toward software subscriptions and recurring revenue.
Nokia and NVIDIA Target a Growing AI-RAN Market
Nokia’s AI and cloud strategy has helped strengthen investor interest in the company throughout 2026.
The AI-RAN launch also came just days before Nokia’s scheduled second-quarter financial results, placing additional attention on the company’s ability to turn AI infrastructure into commercial growth.
Technology research firm Omdia has estimated that the cumulative AI-RAN market opportunity could exceed $200 billion by 2030.
The opportunity is significant, but Nokia must still prove that it can capture a substantial share of that market.
Is Nokia Really the First?
Nokia’s claim that its platform is an industry first requires some clarification.
In June 2026, rival telecommunications equipment company Ericsson began selling its own commercial AI-in-RAN software subscription.
Ericsson said its technology could deliver up to 20 percent higher downlink throughput and up to 10 percent better spectral efficiency. The system has reportedly been deployed across more than 15 live networks.
Ericsson’s solution also runs on existing baseband processors and does not require NVIDIA GPUs.
Nokia’s “industry-first” claim therefore appears to refer specifically to a complete, GPU-accelerated AI-RAN platform rather than the first commercially available use of AI inside a radio access network.
Both companies may technically be correct, but they are competing through different architectures.
Ericsson is adding AI capabilities to existing network hardware, while Nokia is building its platform around GPU-powered computing.
The Risk of NVIDIA Dependence
Nokia’s decision to align its radio network strategy with NVIDIA provides access to one of the world’s most advanced AI computing ecosystems. It also creates a degree of technological dependence.
Nokia Chief Technology Officer Pallavi Mahajan has acknowledged that some parts of the company’s Layer 1 network software are tied to the underlying hardware.
Ericsson has taken a different approach by designing its AI features to remain independent from a specific processor or chip supplier.
Nokia continues to describe its platform as Open RAN-compliant and has highlighted other chip partners, including Marvell. However, the performance improvements being promoted for the new AI-RAN platform currently depend heavily on NVIDIA’s hardware and software stack.
This creates an important contradiction.
Nokia is promoting an open network architecture, but its highest-performance AI-RAN implementation is closely connected to NVIDIA technology.
A Promising Comeback, But Not Yet a Victory
Nokia’s strategy is not without logic.
Partnering with NVIDIA allows the company to rely on an established AI computing leader instead of funding an expensive internal chip development race. The subscription model could also provide Nokia with more predictable revenue than traditional network hardware sales.
Still, the platform is not yet commercially available.
Its most impressive performance claims are targets scheduled for 2027 and 2028, and Ericsson has already reached the commercial market through a different approach.
For Nokia, AI-RAN represents a potentially important comeback for its mobile network business. But the comeback is still being built, tested, and proven.
Its future success will depend not only on AI performance, but also on operator adoption, deployment costs, hardware flexibility, and whether customers are comfortable tying critical network infrastructure closely to NVIDIA’s technology ecosystem.
