The Unbundling of Telecom Has Started

For four decades, the mobile industry has run on one model: own the spectrum, own the customer, own the plan. Bundle those three things, defend each one, and the rest takes care of itself. It worked because there was no other way to deliver mobile connectivity. The vertically integrated MNO was the only model.

That just ended. Not because customers demanded it, not because regulators forced it, and not because MVNOs finally figured out the right pricing. It ended because Starlink became the first “new” global MNO in 40 years. Built from scratch. Same vertically integrated playbook the incumbents used. Just satellites instead of towers. The existence of a genuine alternative carrier is what unbundling has been waiting for, and it's here now.

Mature, vertically integrated industries get pulled apart by software companies that specialize in one layer and serve everyone. The physical layer stays where it was. The customer-facing layer migrates to a software company that abstracts across all the suppliers underneath. It happens in three stages.

First, a software company aggregates supply. It builds a single customer experience on top of multiple infrastructure providers. Expedia did this for hotels. Amazon Marketplace did it for retail. AWS did it for compute. US Mobile is doing it now for telecom: one bill, one app, four networks underneath (three terrestrial plus Starlink), seamless switching the customer doesn't have to think about. Oxio is the platform layer beneath, enabling other brands to build the same kind of product without rebuilding the plumbing themselves.

Second, the customer relationship migrates to the aggregator. Slowly, transaction by transaction. After enough cycles, the customer thinks of themselves as an Expedia user who happens to be staying at a Marriott, not a Marriott loyalist using Expedia. In telecom, this is just starting.

Third, the suppliers commoditize at the layers where competition reaches them. Once the customer relationship lives in the software layer, the infrastructure underneath starts competing on price and reliability to the aggregator, not on brand to the customer. In energy, this happened to power generators first. Companies like NextEra, Vistra, and Constellation now bid megawatt-hours into wholesale auctions every five minutes, while retail electricity providers and solar installers (NRG, Sunrun, Tesla Energy) own the customer relationship. The end customer rarely knows who actually generated their power. But the commoditization wasn't uniform. In states that never deregulated, the integrated utility still owns the customer. And new brands emerged at the orchestration layer that customers do feel loyal to. The pattern in telecom will probably look similar. MNOs become wholesale capacity providers in the most competitive segments, retain customer relationships in others (enterprise, premium consumer), and watch new aggregator brands like US Mobile build the loyalty that used to be theirs. This hasn't happened to MNOs yet, and may not for a decade. But the direction is clear.

Telecom is firmly in stage 1, just entering stage 2.

For a hundred years, electric utilities ran the exact same vertically integrated model telcos have run for forty: own the generation, own the grid, own the customer. Geographic monopolies enforced by physics and regulation.

Then rooftop solar arrived. Tesla, SunPower, Sunrun, Enphase. For the first time, customers could partially substitute for the electricity serviced by the utility. Home batteries extended the electricity used into the night. Electric vehicles added a third layer. Substitutability was partial, not complete. You still needed the grid at night, in winter, in apartments. But many California households now self-generate 60-80% of their annual usage.

Then aggregators emerged to manage the messy in-between. Virtual power plants (Tesla Autobidder, AutoGrid, Sunrun Brightbox, Stem) sit on top of thousands of distributed resources: rooftop panels, home batteries, EVs, smart thermostats. They orchestrate them as if they were a single coordinated entity. The grid operator sees one dispatchable "plant" rather than 50,000 individual homes. The customer sees one app and one bill, without thinking about whether their fridge is running on solar, battery, or grid power. Oxio and US Mobile are building the mobile equivalent: software that aggregates multiple carriers and satellite networks into a single connectivity service the customer experiences as one product.

The utilities split into two groups. NextEra aggressively went into distributed resources and became the most valuable utility in America. PG&E doubled down on the old grid model and went bankrupt twice in a decade. National Grid and Xcel are transitioning toward "platform utilities" that serve as marketplaces for distributed energy. That is precisely T-Mobile's SuperBroadband move, ten years earlier in a different industry.

The deepest lesson: utilities didn't lose by being replaced. They lost by failing to reposition fast enough. Solar didn't have to substitute for grid power in every condition to change the business model. It just had to substitute enough. Starlink doesn't need to replace terrestrial cellular to break the geographic monopoly model. It just needs to be enough of an alternative.

For three decades in mobile, substitutability was weak. Terrestrial carriers all had similar coverage and the same architecture. Dual-SIM devices existed, but managing two numbers and two bills felt like too much work. Most people stayed put as long as they were reasonably happy.Starlink built something different. Not a software layer. Not an MVNO. A complete, vertically integrated, end-to-end mobile carrier: spectrum (satellite plus $17B in EchoStar terrestrial), the radio access network (10,000+ satellites), a cloud-native core, distribution, a consumer brand, eventually a device. The same stack Verizon and T-Mobile built, rebuilt with satellites as the access layer.

But Starlink isn't a finished MNO yet. Spectrum is regulated nationally and needs country-by-country approval. LEO capacity has hard upper bounds; Manhattan at lunchtime would crush the satellite layer. And sub-1GHz frequencies that penetrate concrete walls are held by terrestrial MNOs, so indoor coverage remains hard. Starlink is an alternative carrier, not a replacement carrier. That's the same place rooftop solar was in 2010: a real alternative for some conditions, enough to start unbundling the industry.

The likely resolution is Starlink becoming a hybrid that owns its satellite network but rents terrestrial capacity through MVNO arrangements. The deepest irony: the new entrant that built the most aggressive vertical integration in 40 years will probably end up operating partly as an MVNO too, because even Starlink needs the unbundled stack to deliver a complete consumer product.

The most striking move so far is T-Mobile's launch of SuperBroadband in April 2026. It combines T-Mobile 5G and Starlink into a single managed service: one contract, one bill, intelligent orchestration between networks, coverage in every US ZIP code.

What's remarkable isn't the bundle. It's the pitch. T-Mobile is no longer selling "America's biggest 5G network." They're selling themselves as the company that gives businesses connectivity through whichever network works best, including their competitor's. That's an aggregator pitch, not a network pitch. The largest MNO in America is voluntarily behaving like a software platform sitting on top of multi-source infrastructure they don't fully own.

This is the NextEra move. If the customer relationship is migrating to the orchestration layer, T-Mobile would rather be the orchestrator than let someone else become it.

When the largest MNO in America voluntarily restructures its enterprise product to look like an aggregator, that's the incumbent telling you the old model isn't working anymore. T-Mobile is reading the same structural shift this article describes. That's the strongest evidence for unbundling than any challenger move could provide.

Perhaps none of this would be happening if Starlink hadn't built a parallel carrier. The vertically integrated ambition of one company is what created substitutability across the entire industry, and substitutability is what creates unbundling.Stage 1 has started. Stage 2 is beginning. Stage 3 is a decade away, maybe more, and may never fully arrive if incumbents reposition fast enough. But the trajectory is now visible because we've watched it play out in lodging, retail, compute, and energy. Telecom is following the same path.

Forty years of industry structure is being rewritten right now. The conditions exist for the first time. What happens next depends on the choices every player in the stack makes from here.