What’s the asset of the future, the potential determinant of success for practically everyone in tech? No, it’s not AI. It’s not 5G or 6G or WiFi 7, nor is it fiber or quantum. What is it, then? Real estate.
An astonishing number of critical network and IT assets have one common element—they need to be put somewhere, usually somewhere specific. Increasingly, that means somewhere proximate to points of user connection to networks, because latency is emerging as a key factor in almost all new applications of technology. But even for “mature” things like access networks, the most critical and cost-intensive piece of networks overall, pride of placement is critical. Wireless towers have to be put somewhere wireless users need them, and you have to feed towers with fiber that has to pass along rights of way. FTTH needs the same right of way, and even FWA needs to feed the broadband radio systems as well as site the antennas.
Given all of this, you might think we’ve uncovered the secret asset for telcos to leverage, and twenty years ago or so that might have been true. In the US, for example, we had nearly fifteen thousand telco central offices, and another seven thousand other facilities (parts stores and places for craft personnel to work, for example). We had over a hundred thousand telco-owned cell towers, too. Now? Telcos have been selling off real estate, from towers to facilities, for decades. It’s hard to get the data, but the majority of towers are now supported through shared-lease arrangements, over half the non-CO and almost a third of CO facilities are sold or on the block for 2025/2026. Telcos seem to have been selling off a competitive advantage, but that’s not the whole story.
The big push behind real-estate sales by telcos is due to a combination of shifts in telecom service opportunity and the focus of Wall Street. If you look at a central office, you see something that’s valuable as a point of connection between telecom switches and outside plant, mostly copper loop and remote nodes fed by T/E TDM trunks. Given that there are few places in the world where copper service delivery is profitable or even useful, you can understand how the value of a CO would diminish. Add to that the fact that real estate costs, and selling it not only reduces carrying cost, it generates money to fund things, and it’s not hard to see why there are short-term reasons to sell it off.
Those pressures are often unbearable when you can’t think of any near-term reason not to. We’ve heard for ages how telco real estate could be used, for example, to house cloud data centers, but telco successes in cloud computing aren’t exactly leaping off the financial pages, are they? We’re hearing now that edge computing or GPUaaS are similar ways that telco real estate could be exploited, but if we want to be fair, we have to ask why these “opportunities” are more likely to be exploited successfully than the cloud.
A good question, then, would be whether there’s some other powerhouse player that’ being created as telcos divest their real estate assets. That turns out to be really complicated to determine.
The “real real-estate” portion of assets, meaning actual buildings, seem to have been sold to a wide variety of companies, none of which seems to have any conspicuous technical credentials to flaunt. I’m told, for example, that the largest buyer of non-CO buildings sold by telcos has been trucking companies, perhaps because so many of these buildings were parts warehouses. The one closest to me, for example, was sold to a local trucking firm. The point is that I can’t identify anything suggesting that some secretive tech giants are buying up the buildings.
That may be a tad less true for former central/switching office facilities. While the relatively small number of these that have been sold off doesn’t show a pattern of prospective tech exploitation, some countries seem to be seeing newer acquisitions by companies whose industry affiliation is unclear, or who seem to be real estate investment companies or even shell companies, perhaps proxies for some larger players who don’t want immediate publicity. I think that some of the cloud providers may be arranging for real-estate assets near the edge, just in case.
That qualifier is important. The truth is that there are no current applications that could justify a major edge deployment of IT assets of the type that would validate real estate exploitation. That’s why the telcos haven’t held on to their own assets there. There are, however, two potential applications that could do that, and they have something in common—a direct link to the real world. In fact, that’s the only thing that can make the edge truly valuable in the first place.
Real-world applications are those that require synchronization with real-world elements, which means they have to be run in real time, with minimal latency between the application and the elements and processes they’re involved with. They also have to be highly reliable, because controlling a real-world process means creating a potential real-world disaster if you mess up. I’ve blogged about both in the past, so I’ll just mention them here.
The first is the empowerment of the portion of the workforce that’s not desk-bound, which makes up roughly 40% of all workers and slightly more than that portion of the total unit value of labor. We’re talking some steps now to realize the benefits of this empowerment, but only by slightly expanding the scope of IoT applications relating to industrial and manufacturing processes to cover multiple facilities and the interconnecting transportation resources. The applications are still largely local in nature, meaning they can be addressed with on-premises hosting, but if the scope of operation gets larger and the interactions more complex, cloud/edge hosting is likely required.
The second potential application is truly transformational, enough to touch everyone’s life. Fill the real world with sensors. Give each person an “assistant/agent”, likely based on AI, and give every point of purchase and service the same thing. Drawing on the mesh of sensors, and the interests of both buyers and sellers, AI agents would interact to help everyone work and live, finding thing we want, helping us with our tasks, even entertaining us.
Most of the technology needed to support either of these applications is already available, and what remains is more a matter of policy than of technology. If tech is to do more for us, it has to integrate more tightly with our lives and work, with our world. That raises issues about privacy and security that we’ve been, so far, reluctant to address. Can an application that helps us work also monitor how hard we’re working, and can one that can help us find each other or find a product or service we want to buy, also stalk us?
Hype may save us here. I think the overwhelming amount of hype we deal with is largely due to the fact that, absent things of real value and interest, things are invented or exaggerated. Might these two applications benefit either by becoming hype targets, or by becoming a way to realize the exaggerated value of things like AI. We can hope, not so much for next year but perhaps soon thereafter. Maybe even for the telcos, while they still have real estate to play in the new games.
Happy New Year.
