Summary
Nokia has launched what it calls the industry's first AI-native radio access network (RAN) platform, built with NVIDIA's technology. The platform aims to help mobile operators get more capacity from their existing spectrum without buying new hardware. This launch is a key part of Nokia's plan to turn around its struggling radio business, but it faces competition from Ericsson, which already has a similar product on the market.
Main Impact
Nokia's new AI-RAN platform, announced on July 15, 2026, marks a major shift in how the company builds radio networks. Instead of relying on its own custom chips, Nokia is now using NVIDIA's powerful GPUs and software. This move is designed to cut costs and speed up innovation. For mobile operators, the platform promises big gains in network performance without the need for expensive hardware upgrades. However, the platform is not yet available for purchase, and its most impressive performance targets are still years away.
Key Details
What Happened
Nokia introduced its AI-RAN platform, which combines its anyRAN software with NVIDIA's Aerial computing system. The platform uses artificial intelligence to make radio networks more efficient. Operators can buy it as a software subscription, not a hardware box. They have three choices: a plug-in card for existing Nokia sites, a standalone AI-RAN node, or a cloud-based version from partners.
Important Numbers and Facts
Nokia says the platform has already shown over 20% gains in spectral efficiency – how much data can be sent over a given slice of radio spectrum. The company aims for 50% gains by 2027 and more than 100% by 2028. That would mean operators could roughly double the capacity of their current spectrum. But these are targets, not proven results. Pilot tests are expected by the end of 2026, with commercial sales starting in 2027.
Background and Context
Nokia's radio business has been struggling for years. CEO Justin Hotard, who took over in 2025, has called it the company's hardest problem. In late 2025, he told investors the mobile business was not making enough money. The partnership with NVIDIA, which included a $1 billion investment from the chipmaker for about a 3% stake in Nokia, is central to fixing this. By using NVIDIA's chips and software, Nokia can spend less on its own chip development and focus more on software. This is a big change from its old model of selling custom hardware.
Public or Industry Reaction
Investors have reacted positively. Nokia's stock price has gone up through 2026, helped by its AI and cloud plans. Research firm Omdia estimates the total AI-RAN market could be worth over $200 billion by 2030. But the "industry's first" claim is being questioned. Ericsson started selling its own AI-in-RAN software in June 2026, and it works on existing hardware without needing new GPUs. Ericsson says its product delivers up to 20% higher downlink speeds and up to 10% better spectral efficiency, and it is already in use on more than 15 live networks. Nokia's claim to be first is based on being the first to use GPU acceleration for AI in the RAN, which is a different approach.
What This Means Going Forward
Nokia's strategy is a bet on NVIDIA. The company's AI-RAN platform is tied to NVIDIA's technology, and its performance gains currently depend on it. This creates a risk if NVIDIA's chips become too expensive or if a better alternative appears. Ericsson, on the other hand, has chosen to keep its AI features independent of any single chipmaker. Nokia says its platform is Open RAN-compliant and works with other chips, but the best performance comes from NVIDIA. For now, Nokia's comeback is a work in progress. The platform is not yet commercial, its biggest promises are for 2027 and 2028, and a rival is already selling a similar product. Nokia's future in radio now depends heavily on how well it can execute this plan with NVIDIA.
Final Take
Nokia's AI-RAN platform is a bold move to save its radio business by partnering with the leader in AI chips. It offers a clear path to better network performance and a new subscription revenue model. But it is not a sure thing. The technology is not yet proven at scale, the competition is already in the market, and Nokia's success is now tied to NVIDIA's roadmap. This is a comeback story that is still being written.
Frequently Asked Questions
What is an AI-RAN platform?
An AI-RAN platform uses artificial intelligence to make radio access networks – the part of a mobile network that connects phones to the internet – work better. It can improve data speeds, reduce delays, and get more use out of the radio spectrum a mobile operator already owns.
How is Nokia's platform different from Ericsson's?
Nokia's platform uses NVIDIA's GPUs to run AI tasks, which it says gives better performance. Ericsson's platform runs on existing baseband chips and does not need new hardware. Ericsson's product is already on the market, while Nokia's will be available in 2027.
Will this make my mobile internet faster?
If mobile operators adopt this technology, it could lead to faster and more reliable connections. The platform is designed to increase network capacity, which means less congestion and better speeds, especially in busy areas. But it will take time for operators to buy and install the new systems.
.jpg)