When someone on LinkedIn posted about Nokia’s launch of AI-RAN, I commented that I was sorry Nokia was drinking the AI kool-aid. Of the 88 operators I chat with, 64 reacted to the launch and my comment, and I think the story they tell is interesting.
My view here is simple. Nokia wants to sell wireless infrastructure to operators, and a major 3GPP standard update is certainly a historical driver of major investment. The problem is that 5G didn’t deliver on the promises made for it, that there would be compensatory revenue generated to offset costs. When 6G work started, operators said they didn’t want “fork-lift” upgrades because they feared that the business case hype that surrounded 5G would likewise infect 6G. Fork-lift upgrades are what vendors want, of course, and if you look at the proposed 6G targets, they sure sound like the same stuff 5G was supposed to deliver and didn’t. So, what’s Nokia to do? Jump on the AI hype bandwagon.
Of the 64 operators who commented, all agree that their company uses AI. In fact, all said they used cloud-hosted AI, and 53 said they were totally blowing their budgets on the services. Those of my contacts who are truly AI literate say that their cloud AI costs far exceed what it would cost to deploy their own resources.
However, only 17 said they actually hosted any AI on the premises, and only 11 said their deployment was targeting generalized applications; the rest said that they had AI “embedded” in some application. For Nokia, the good news was that all 64 believed that AI-RAN could “enhance mobile network capacity and efficiency”, to quote the article I cited above. The bad news was that only 6 thought that AI-RAN gains would generate enough to actually make a business case for massive 6G spending.
The reason I picked the article I cited, which is a Wall-Street-focused analysis of Nokia, is that it demonstrates the issue that network vendors, including Nokia, face. The company is over-valued, says the article, and I agree. The reason is that you can’t keep increasing spending on tech when your own revenues can’t keep pace with the increase, or you look worse on the Street, which no public company can afford to do for very long. Network spending, across both enterprises and network operators, faces the problem of having no new revenue to justify major new spending plans.
I always told my consulting clients that a problem of public perception was a PR problem, one that had to be solved through marketing messaging because “Bulls**t has no inertia”. So, like many other companies, Nokia is jumping on the AI hype wave because it will provide them some air cover for what Street analysis says is an excessive valuation.
Getting cover, of course, isn’t solving the problem. Of my 64 operators, all agree that the only long-term solution to their current challenges is new revenue. Of the 64, 49 think that has to come, eventually, from higher-layer services. The remaining 15 still believe that new connectivity revenue is possible, but all 64 agree that it is likely of not certain that their initial revenue gains would have to come from connectivity. Where is this new connectivity coming from? Again, all 64 think it has to come from something related to IoT.
Pretty much all the telco infrastructure companies have touted IoT in the past, but the problem has been a combination of pathetically naive positioning (sell to machines because you aren’t limited by birth rate and maturation) and insipid offerings (device registration services without any notion of why the devices would be needed). Nokia actually did a bit more to expose things like the dependence of 6G on private 5G and the applications of private 5/6G than most vendors. However, they didn’t address the fundamental problem of new services, new benefits.
Which is? Well, if you look at networking today, you see applications that existed and were then connected. Connectivity only required that you prove marginal utility gains by extending application scope and capabilities via connectivity. Today, though, we’ve largely run out of applications to exploit. Transaction processing is the core of business networking, and entertainment the core of consumer networking. Been there, done those, got the tee-shirts. To open new connectivity opportunities today, we need new applications, which means that the infrastructure and willingness to pay for those has to be worked through. Opportunity in network services, in 2026 and beyond, needs an ecosystem of players to develop, and telcos aren’t much good at even thinking about ecosystemic issues.
Nor, perhaps, are their vendors. However, some of my friends in Nokia (who agree that AI-RAN isn’t going to make a 6G business case) think that the missing step is that the edge deployment of AI-RAN results in an edge deployment of AI, which in turn gives the telco first an opportunity to sell edge hosting, and then to rise up from an IaaS model toward an SaaS mode, meaning to host AI-based tools to facilitate IoT missions. Most of these people, though, are not confident that Nokia is going to exploit that transition of mission, or even that senior management recognizes what would be needed. The great majority agree that they’re not doing the necessary stuff today.
What’s needed today is an ecosystem-builder, and I think that the initial focus has to be on the use of private cellular technology and the “facility connectivity” mission. The basic problem with building an IoT ecosystem for 6G to support is that if 6G is needed for that ecosystem, you can’t start it without having telcos bear that dreaded “first cost”, the period when investment is huge and returns are yet to develop. But spread IoT out from an assembly line to a whole multi-structure facility doesn’t require 6G service, only facility-wide connectivity. You could do it with private 6G, even 5G. In verticals where a “company” is made up of multiple “facilities” that are metro-concentrated, you can bridge your facility starting point with targeted WAN connectivity.
To make this work, Nokia and other vendors would need to collect the players who would contribute value and gain returns, a broad enough group to assemble solutions for the target verticals. And they’d have to do this in a very public way, to stimulate coverage and create some air cover, even tie in with AI, which is itself in great need of some practicality injection. Could Nokia do this? I think they could. Will they? That’s a harder question, one even my Nokia friends can’t answer confidently.
Nor can the telcos, which is a shame given that they have the most to gain from an answer.
