Nokia has thrown its weight behind a new open RAN interface called E3, aiming to support "dApps" — network applications that execute faster than either xApps or rApps — as it tries to make its 5G portfolio more competitive against Ericsson and Samsung.

The move matters because Nokia's survival as a mobile kit vendor deep into the 6G era depends on large western telcos embracing its argument that Nvidia's GPUs are both a desirable and economical feature of the radio access network. Nokia's latest custom silicon product, co-developed with Marvell Technology, is the last it plans to sell. Within a few years, there will be no alternative to Nvidia.

That wager on Nvidia GPUs helps explain Nokia's eagerness to publicize telco interest in its forthcoming AI-RAN offers. After its recent Radio World event in Oulu, Finland, the company said operators are testing the technology, with 19 having signed up for "initial deployments," according to CEO Justin Hotard in a LinkedIn post. Yet the three companies he identified — Pulsone (a brand of Cohere Technologies), DeepSig and ZaiNar — are startups designing RAN software, not telcos.

Still, Nokia has looked the most willing among the big kit vendors to let other developers inside its networks. That openness was baked into the AI-RAN pitch that came from Nvidia, whose Aerial framework for a RAN software stack runs on the CUDA platform and is open to developers who know how to work with it.

"It provides a base for our teams and Nvidia to go out and test some of the models," Pallavi Mahajan, Nokia's chief technology officer, told Light Reading in the summer. "I think having that as a base is super useful, which is why we don't look at Nvidia as just a silicon provider."

Nokia built on its earlier openness to the O-RAN Alliance movement, which standardized the fronthaul interface joining radios to baseband units and is developing specifications for hosting third-party network apps. Those apps come in several flavors, but only one — the rApp, sitting on a non-real-time RAN intelligent controller — has been wholeheartedly adopted. Ericsson, Nokia and Samsung all support rApps for tasks that do not require low-millisecond execution speeds.

The xApp is a different story. It requires a near-real-time RIC and the E2 interface, and neither Ericsson nor Samsung has sounded supportive, judging by earlier public statements. Both appear to believe xApp-type capabilities are best delivered through their own baseband products. Nokia flagged E2 trials in 2023 but made no observable progress afterward, as it acknowledged at a press event in Oulu in late 2025. "xApps has been of less interest lately," said Tommi Uitto, the former president of Nokia's mobile networks business group.

James Crawshaw, a principal analyst with Omdia, was similarly downbeat. "Ericsson doesn't advocate xApps and is only pushing the rApp story," he said. "Moving to xApps, where you're changing things at very, very fast control loops — that's a future project."

Enter the dApp, presented by parts of the industry as an option for tasks demanding even faster execution via a real-time RIC. Cohere, one of the startups named by Hotard, has designed a dApp version of its capacity-boosting universal spectrum multiplier (USM), having previously built an xApp for the same product. Where xApps needed E2, dApps would use E3, which Nokia now supports.

"We've taken the whole concept of APIs and decoupling layers in a very serious fashion, and what it now enables us to do is bring in an O-RAN-compliant E3 interface," said Mahajan. That will let Nokia support apps for sensing, more "geo-precise" location monitoring and improved spectral efficiency, she explained.

A Cohere partnership could make Nokia's target of doubling spectral efficiency by 2028 on much of the existing infrastructure look more achievable. Vodafone, Telstra of Australia and Bell Canada have publicly endorsed Cohere's USM technology after field trials, and all three have invested in the startup. Cohere CEO Ray Dolan identified Nokia as the vendor partner during the Bell Canada trials, which used an off-the-shelf server deployed at mast sites. A dApp would represent a closer, direct integration.

As Cohere explained in a July blog post, E2 operates primarily within the control plane, while "the E3 interface extends this model into the user-plane domain, allowing dApps to influence scheduling, precoding, and channel estimation with sub-millisecond latency."

Commercial reality lags the concept, however. "rApps were easy, xApps were hard, dApps are really hard," Dolan acknowledged. "It's a great idea that is nothing but a great idea." The engineering is as "complicated as brain surgery," he added, and the industry has substantial work left to do.

Ericsson, the world's biggest non-Chinese 5G vendor, is keeping its distance. "We view dApps as an emerging concept, not a replacement for rApps," the company said by email. "rApps remain our established approach for automation use cases within our architecture." For xApps and dApps, it said, "we continue to monitor and research, but there is no product commitment to support dApps today."

The divergent strategies also raise questions about app portability. Nokia is building an ecosystem around Aerial, but CUDA is native to Nvidia GPUs, meaning apps based on it cannot migrate to other hardware — and neither Ericsson nor Samsung is actively pursuing a GPU RAN strategy today.

None of this addresses the problem open RAN was meant to fix: the shortage of economically viable suppliers building the main RAN parts. Operators remain dependent on a handful of big vendors, with market spending falling from $45 billion in 2022 to $35 billion last year, according to Omdia. The analyst firm predicts rApps will generate annual revenues of only several hundred million dollars by the end of the decade.

If Nokia's dApp-led approach gains traction — and reverses the market share losses of recent years, notably in the lucrative US market — competitors may be forced to respond. For now, with many telcos seeing little need to replace current hardware, the burden of proof rests on Nokia as it works toward its 2028 spectral efficiency target.