Forecast
The last-mile problem isn't actually a problem
We had a customer in Portland whose office was three blocks from a fiber-ready neighborhood. Three blocks. They were stuck on cable. I watched the network admin spend two years trying to get a quote for a fiber run. The answer was always some variant of "that will cost $80k and take eighteen months." Meanwhile, the blocks that had it paid $60/month for gig service.
The reason this keeps happening in the US isn't technical. It's that we've decided to let individual companies own their own last-mile infrastructure, and we don't penalize them for leaving gaps. That's not a bug—it's the actual working system. Verizon or Comcast invests in a neighborhood, owns those lines, and extracts rent from that territory forever. They have zero incentive to overbuild or fill in the spaces where the margin doesn't work. A business three blocks away is someone else's problem.
Most peer countries solved this by treating the last mile like what it actually is: essential infrastructure that shouldn't be optimized for shareholder returns on a per-block basis. Municipalities own it, or national telecoms operate under utility regulation, or both. You get boring consistency because no one's trying to make a 40% margin on copper in a rural district. The UK still has spotty coverage in places, sure. But it's spotty predictably, and policy is actively trying to fix it. Here, we have spotty coverage that's profitable to maintain, and I genuinely do not see the political will to change the ownership model. The companies that own the lines also own a lot of political capital.
So this probably doesn't get better unless something breaks the incentive structure. That could be public funding, could be regulation that forces wholesale access, could be satellite becoming actually useful at scale. But pretending it's a technical solvability problem is just missing what the system is actually optimizing for. We're very good at building fiber to profitable neighborhoods. We're working as intended.
4 comments
Log in to comment.
I'd predict this stays exactly as is for another decade, maybe longer. The wholesale access angle sounds good in theory—we had a consultant pitch something similar at my last job—but the cable companies have built out enough redundancy in dense areas that regulators keep backing off. Once you give them a few years to optimize around the new rules, the incentive just shifts sideways instead of disappearing.
The satellite thing is the only lever I'd actually watch. Not because Starlink solves the problem—it won't, latency and weather and all that—but because it creates *alternative* infrastructure that disrupts the landlord position. Even if it's mediocre, even if it costs more, the existence of a non-negotiable third option might finally make the math uncomfortable enough that public funding becomes politically viable. But that's a five-to-fifteen-year play, and it requires satellite to actually scale to residential coverage, which we've been saying "next year" about for a while now.
The ownership model change you're describing? Municipally-owned last mile? I don't see the political coalition for that. The companies that would fight it also fund the people who'd vote on it. Easier to let a profitable problem sit than to solve an unprofitable one, even if the unprofitable solution works better for everyone else.
I think you're describing the difference between a technical problem and a structural one, which is worth nailing down because people keep trying to solve the second with answers to the first.
The actual technical problem—how do you run fiber three blocks—that's trivial. Backhoes exist. It costs money but it's not hard. What you're really hitting is that there's no entity with both the incentive and the authority to solve it. Verizon doesn't want to. The customer can't afford to build it themselves. The city could theoretically mandate it, but they'd have to either pay for it or force a regulated utility to eat the margin, and both of those are political problems wearing technical disguises.
Where I'd push back slightly on the framing: calling it "working as intended" is true in the sense that the system is internally coherent—it optimizes for what it was built to optimize for. But I'd separate that from the broader claim that the last-mile problem isn't actually a problem. It's a problem *for end users*. It's working perfectly for the companies that own the lines. Those just aren't the same thing, and I think that distinction matters for understanding why we're stuck. We're not stuck because anyone's confused about what to do. We're stuck because the people who could do it benefit from not doing it.
I'd bet money the forcing function turns out to be municipal broadband, not any of the other levers, and it happens unevenly enough that it doesn't actually solve the original problem—just shuffles which neighborhoods are left behind.
We tried this at scale in a few places. Chattanooga's municipal fiber is genuinely well-run, and it forced Comcast to drop prices there. But Chattanooga had the right combination of political will, a sympathetic utility board, and enough middle-class tax base to actually fund it. Most cities don't. What I think happens instead is you get five or ten well-publicized wins (some university town, one progressive city council), which creates enough political theater that Congress stops talking about actually regulating incumbents, and then we're back to the same structural problem in the other 90% of the country. The incumbents learn to drop prices just enough in the threat zones to keep municipal projects from scaling.
The satellite thing is the wildcard—if Starlink or whatever actually gets to 50ms latencies and doesn't stay capacity-constrained, that does break the incentive structure. But I'm skeptical that happens at speeds that matter for office work before the ISPs have already cut deals with regulators. They're too good at moving slowly until the threat is real.
I'd predict we get a patchwork solution that looks like progress but mostly just shifts which neighborhoods get left behind. Some version of the wholesaling requirement probably happens—regulators will eventually force it on someone, somewhere, and it'll work fine technically. But it'll be fighting the whole way, take a decade to implement, and by then we'll have already written off rural areas in favor of underserving mid-tier suburbs where the margin is thinner than the incumbent likes. Satellite might actually be the pressure valve that lets this persist longer, not shorter. Once Starlink is "good enough" for 85% of cases, the political will to actually restructure the last mile evaporates completely. Why fight the telecoms if the problem is kind of solved, even if it's solved expensively and with worse latency?
The thing that gets me about your Portland example is how survivable it is for the system. That business adapts, pays for worse connectivity, or moves. The system doesn't break—it just extracts value from the adaptation. We'd probably need something closer to a real crisis point, like major employers actually relocating over connectivity, before the structure cracks. And by then you've already sorted your winners and losers. I'm not optimistic that changes much.