Nvidia and Nokia Expand AI-RAN Work as Carriers Test Paths Toward 6G

Nvidia and Nokia are developing commercial AI-RAN products that combine accelerated computing with radio-access network software. T-Mobile and other carriers are participating in tests, but the work remains an early step toward 6G standards and commercial networks rather than a finished 6G service.

At the center of the announcement is a new piece of hardware NVIDIA is calling the Arc Aerial RAN Computer, a telecom computing platform the company describes as 6G-ready. Nokia plans to fold NVIDIA-powered, AI-RAN products into its existing radio access network portfolio, while Dell Technologies is contributing PowerEdge servers to help power the resulting systems. It's a genuinely unusual lineup, chipmaker, telecom equipment vendor, server maker, and wireless carrier all working from the same architecture at once.

Why networks are using AI

The pitch behind AI-RAN is straightforward, even if the engineering underneath it is not: today's cellular networks run on radio access algorithms that are, in NVIDIA's own description, roughly fifty years old at their core. Threading AI directly into that layer, rather than treating it as an add-on service running elsewhere, is supposed to squeeze meaningfully more spectral efficiency out of the same physical airwaves, while also giving networks the horsepower to handle a coming wave of AI-heavy traffic, from generative AI assistants on phones to autonomous drones and robots that will need near-instant, reliable connections.

NVIDIA's telecom lead, Ronnie Vasishta, has described this as building the intelligent network infrastructure that what the industry calls physical AI, autonomous machines actually operating in the real world, will ultimately depend on. It's a bigger ambition than simply making phone calls clearer or downloads faster; the argument is that the whole next wave of AI-powered hardware, from self-driving cars to factory robots, needs a wireless network smart enough to keep up with it in real time.

Competition and geopolitical considerations

NVIDIA isn't approaching this alone or quietly. The company has built out an AI-RAN Alliance that now counts more than 130 participating companies, and has separately joined the US FutureG Office's OCUDU Initiative, a government-industry effort aimed at accelerating open, software-defined, AI-native 6G architecture domestically. NVIDIA has also put real money behind the bet, investing roughly $1 billion in Nokia as part of its broader push to use 6G as an entry point into the telecom hardware business it's historically had little presence in.

That flurry of alliances and investment isn't happening in a vacuum. Industry analysts tracking the space note that geopolitics has made the race toward 6G considerably more interesting than prior generational shifts, with multiple countries racing to shape the technical standards and supply chains that will define the next era of global wireless infrastructure. For the US specifically, backing an AI-native approach to 6G is being framed by some in the industry as a chance to reclaim ground in wireless leadership that shifted overseas during earlier network generations.

Technical and commercial uncertainties

For all the momentum, plenty of skepticism remains about how fast, and how completely, this vision actually gets built. Formal 3GPP standards work covering dedicated 6G studies isn't expected to wrap up until the 2026-2028 window, with commercial specifications not fully codified until closer to 2030 or 2031. Even within the industry, there's genuine debate about just how far AI hardware should physically extend into network infrastructure. A Samsung networks executive has noted that operators are still carefully weighing the total cost of actually putting GPUs inside base stations against the business case for doing so, a sign that the economics here are far from settled even as the technical vision gets more concrete.

Market forecasters, meanwhile, are penciling in a long runway before this translates into real revenue. Global technology intelligence firm ABI Research projects total 6G radio revenue will grow from roughly $8.1 billion in 2029 to $67 billion by 2034, with AI-chipset radio revenue specifically climbing from $1.8 billion to nearly $58 billion over that same stretch. Researchers at the firm describe 2029 through 2031 as the critical proving period, when operators will find out whether enterprise demand and AI-driven services can actually justify the scale of investment being discussed today.