So far, 2025 hasn’t been a happy year for enterprise network professionals, from what they’re telling me. They have more budget pressure than at any time in their careers, and at the same time they have no clarity on their own company’s overall business strategy. Every technology they read about seems over-hyped, their familiar vendors are doing unfamiliar things, and many face a risk of layoff for the first time in their careers. Even their friends who work for vendors or startups are uncertain.
OK, time to think about a side hustle or new career? No, that means it’s time to get to work, and the good news is that some enterprises are already leading the way by exploiting the reality behind the great hype waves that have disappointed so many. The hyped technologies are private 5G, edge computing, and AI, and what’s important about them is that they need to be considered together, symbiotically. I can offer limited details on two enterprises that have demonstrated that symbiosis and can show us what the future might look like.
Most enterprises have absolutely no need for private 5G, and most know that now. Some do, and the private 5G value proposition in those cases is not only solid but expanding geographically and along the supply chain.
One manufacturing enterprise that has some factories that use a highly agile and adaptable manufacturing system (including computer-controlled machines, robot-like transportation, and automated part and goods movement) switched from a combination of wired IoT and Wi-Fi to private 5G in one facility. They change reduced their control-loop latency overall to a mere one-fifth of what it had been, which let them speed their processes by more than 90%. They’ve started adding other facilities now, and they’ve expanded the private 5G to nearby sites to move parts more efficiently to the assembly area and carry goods away. They’re now talking with a cloud provider about hosting a private web of 5G over an area of almost a thousand square miles to allow for computer control of all goods and parts movement and assembly processes. They say they’re going to explore augmenting this with public 5G services next year.
The computing facilities used in this first example are co-located with the manufacturing operations, but that’s not the case with the second example.
This second enterprise thinks they can convince an operator or a cloud provider to expand on a private 5G experiment they ran using an Open RAN implementation. Open RAN includes what’s called a “RAN Interface Controller” (RIC), and it manages hosting for what the specifications call “Near-Real-Time” and “Non-Real-Time” components of RAN logic. The enterprise used the Near-RT-RIC as an edge-computing component for an IoT application, and the Non-RT-RIC as a link to their cloud computing resources. The experiment has delivered acceptably low network and compute latencies without co-locating compute facilities and processes, and the company wants to find cloud or mobile operator partners to expand this to a continental level. They’re already in discussions, they say.
This brings us to AI. In both these enterprise examples, we see the concept of a new technology model deploying from a seedling-like start in a single location and then expanding outward and, at the same time, expanding to other related areas of business operation. Through this double-build-out, to be effective, the companies have to sustain the pace of operations in each facility, which means that effectively, 5G is creating a big, single, facility, each piece of which has to synchronize with the other. To make that happen, guess what’s used? AI.
Enterprises are building great AI applications, just not the sort we usually hear about. AI isn’t some Yoda-like genius sitting on the shoulder of every worker in a successful deployment, it’s a part of a business application workflow, a software component. For every one of those Yoda-like AI applications that make a minimal ten percent ROI, enterprises have found almost five workflow deployments that can double that, making this sort of AI the only kind that makes aggressive business cases. Yoda works through the worker; the optimal AI isn’t limited by human actions. It does contained things, complex things, very quickly. That’s why it’s an essential piece of both these enterprise examples I’ve talked about.
Think about a supply chain like those old-fashioned bucket brigades, every step depending on being synchronized with the ones before and after. Then think about how hard it would be if you were moving things of different sizes, requiring different holds, tools, even gloves. To make that work efficiently, you’d have to be able to anticipate what was needed at the end of the line and warn everyone upstream to start prepping in time for the shift, right? That’s where AI comes in. If a plant is going to shift between products at a given time, AI can time the shift in the parts supply upstream along all the paths impacted, and the delivery processes downstream as well. With a boost from AI, not only can all these ancillary manufacturing, transport, and warehousing steps be coordinated, but also parts and materials sales orders can be timed and product availability updated for sales.
Two examples don’t make a trend, but they may mark a start. Here we have examples of really exploiting three technologies that are popularly, and accurately, seen as over-hyped. Why then the negative characterization of these three? The answer is that, taken individually, they are over-hyped. There are no single-technology-low-apple targets left out there to aim at. Everything is like a multiple-bank-off-this-ball-onto-that sort of pool shot. We can’t address the next step in computing or networking without a symbiosis of key technology advances. We have the advances, but not the symbiosis.
Except here, in these examples. Why? Because they’re locally initiated, they start small, they contain investment and risk. And, of course, because they can expand to fill the necessary business space as they prove out, without risky technology shifts. That’s why private 5G might be the most important technology for enterprise innovation of our age. Not because it’s private, but because it’s 5G, the stuff building mobile infrastructure worldwide. Private 5G is 5G in a sandbox, a place to play with application approaches that can become an entire landscape when it’s ready.
