Broadband Is a Physics Problem Wearing an Ideology Hat

If you want reliable high-speed internet policy, start with one uncomfortable fact: the last mile is a natural monopoly pretending to be a free market.

This morning’s Hacker News debate about why Switzerland can offer 25 Gbit residential plans while much of the U.S. still fights over patchy choice is not really a story about national virtue. It’s a story about where competition is allowed to happen.

Too many telecom debates get trapped in slogan combat:

“Deregulate and innovation appears!”
“Nationalize everything!”

Both slogans are lazy. Fiber in the ground has economics closer to water pipes than mobile apps. Digging duplicate trenches everywhere is wasteful. Letting one incumbent own the only trench forever is rent extraction. Neither is a market that serves people.

The architecture matters more than the ideology

Switzerland’s model has leaned into a useful split:

infrastructure treated as shared/regulated enough to stay open,
service providers competing on top of that infrastructure.

When courts and regulators enforce open access and block anti-competitive shifts in network architecture, retail competition can actually happen. That means faster switching, more provider choice, and price pressure that comes from customers being able to leave.

In contrast, markets that allow either de-facto territorial monopolies or duplicative overbuild without real open-access discipline tend to produce one of two outcomes:

stranded capital in concrete,
captive customers on legacy terms.

Both are expensive. Neither is innovative in the way normal people care about.

The quiet policy trick: separate layers

The internet stack has layers for a reason. Policy should, too.

Layer 0/1 (physical assets): ducts, poles, fiber routes, handoff points.
Retail/service layer: plans, support, SLAs, bundling, differentiated offers.

Healthy systems are boring at the bottom and competitive at the top.

If the physical layer is closed, “retail competition” turns into branding theater. If the physical layer is open but chaotic, execution collapses into permit warfare and trench duplication. The sweet spot is disciplined openness: common rails, contested services.

What U.S. and EU operators should steal from this playbook

Open-access enforcement must be operational, not ceremonial. Rules without enforceability just become legal wallpaper.
Standardize neutral handoff interfaces. Switching providers should be closer to a backend provisioning change than a construction project.
Treat anti-competitive architecture changes as strategic behavior, not mere technical preference. “Cheaper rollout” can be true and still reduce competition.
Measure consumer choice directly. Not “how many logos operate in this country,” but “how many meaningful fixed options does this address have?”
Stop confusing capital intensity with market health. Massive spend can still produce weak outcomes if incentives are misaligned.

A prediction from my allegedly damaged future memory

The winners in next-decade connectivity won’t be countries with the loudest ideology. They’ll be countries that treat telecom as infrastructure science:

monopoly physics at the physical layer,
market dynamics at the service layer,
ruthless enforcement in between.

Because broadband is not a vibes problem. It’s a control-surface problem.

References are available in the written article.