The Political Economy of Fast Internet: What the US Gets Wrong About Infrastructure

A Swiss ISP called init7 sells a product called Fiber7-X2 that delivers 25 Gbit symmetric fiber to residential customers. Not a business plan. Not a special municipal contract. A consumer internet subscription. Meanwhile, the US Federal Communications Commission spent years debating whether 25 Mbps down and 3 Mbps up even qualified as broadband, a definition so outdated it described connections slower than a single 4K video stream.

This gap exists because of policy, not physics.

What Switzerland Actually Built

Switzerland’s high-speed internet story is partly about init7 and partly about the underlying infrastructure they ride on. The Swiss fiber buildout happened through a combination of Swisscom (the partially state-owned incumbent), municipal utilities, and regional cooperatives. These entities laid fiber to the premises — real FTTH, not fiber-to-the-cabinet with a copper tail — and then opened that infrastructure to competing ISPs under regulated wholesale access rules.

This is the model that makes init7 possible. They do not own the fiber in the ground. They lease access to it at regulated rates, then compete on service, price, and speed tiers. init7’s Fiber7 product line starts at 1 Gbit and goes to 25 Gbit symmetric. Their pricing is transparent and published. When sschueller wrote about this, the core point was that Switzerland achieved this not through some magical free market dynamic, but through deliberate infrastructure policy followed by competitive access on top of it.

Swisscom is roughly 51% owned by the Swiss Confederation. The government did not build fiber because it was profitable — it built fiber because telecommunications infrastructure is a public good with natural monopoly characteristics, and waiting for the market to do it would mean waiting forever.

The Natural Monopoly Problem the US Never Solved

Broadband infrastructure is a natural monopoly in the same way that water pipes and electricity grids are. The marginal cost of adding a second subscriber to an existing network is low. The fixed cost of building a competing network to serve the same area is enormous. This means that in most residential markets, especially outside dense urban cores, it is economically irrational to build two competing fiber networks.

The US Telecommunications Act of 1996 was supposed to solve this by requiring incumbents to lease their infrastructure to competitors at regulated rates, similar to what Switzerland does with fiber. It worked reasonably well for DSL in the late 1990s, which is why there was briefly a competitive ISP market. Then the FCC, under pressure from incumbent carriers, progressively narrowed the definition of which infrastructure had to be shared, and by the mid-2000s the sharing requirements had been largely gutted for the new broadband infrastructure being deployed.

The result was predictable: incumbents deployed broadband on their own terms, in the areas where it was profitable, at the speeds they chose, with no competitive pressure to do better. By 2024, roughly 19 million Americans still lacked access to fixed broadband at all, and tens of millions more had only one provider available to them.

A market with one supplier is not a market. It is a monopoly. Calling it a free market because the government is not the supplier does not change the structural reality.

How Incumbents Captured the Regulatory Process

The more precise story is not that the US chose free markets over regulation. The US chose a particular form of regulation: one that protected incumbent infrastructure owners from competition while allowing them to set prices and speeds without oversight.

Municipal broadband is the clearest illustration of this dynamic. When cities like Chattanooga, Tennessee, or Wilson, North Carolina, decided to build their own fiber networks because private ISPs would not, the incumbent carriers lobbied state legislatures to pass laws prohibiting municipal broadband. As of 2024, 17 states had such laws on the books. This is not deregulation. It is regulation explicitly designed to prevent competition from emerging.

Chattanooga’s EPB fiber network, launched in 2010, delivered 1 Gbit symmetric service years before Comcast or AT&T offered anything comparable in most of their service areas. It was built by a municipal electric utility that already had infrastructure in the ground. The network has since been credited with significant economic development in the region. The incumbent carriers fought it in court and in the legislature at every stage.

This is the pattern the article describes as the free market lie: the claim that slow, expensive internet in the US reflects the natural outcome of market forces, when the actual outcome was engineered through regulatory capture and anti-competitive legislation.

What the Fast Countries Have in Common

Switzerland is not an outlier. South Korea, Japan, Singapore, and the Netherlands consistently rank near the top of global broadband speed charts. They share several characteristics that the US does not.

First, they treated broadband infrastructure as infrastructure, not as a product. Governments either built the physical layer directly or heavily subsidized it through public utilities, then allowed competition at the service layer. South Korea’s Informatization Promotion Fund funded a national fiber rollout in the early 2000s. Japan’s NTT, a privatized but government-origin carrier, built fiber to most of the country under regulatory requirements to open access.

Second, they mandated open access on that infrastructure. Any ISP can offer service over the physical layer. This creates actual competition on price, speed, and service quality, rather than competition between two cable companies that have informally divided territory.

Third, their regulatory bodies maintained clear definitions of broadband that tracked actual usage patterns rather than political compromise. The European Union’s 2025 broadband targets include 1 Gbit connectivity to all households and 10 Gbit to socioeconomic drivers. The US FCC only updated its definition of broadband to 100/20 Mbps in 2024, a figure most European countries would consider a baseline minimum.

The Infrastructure Investment and Jobs Act: Late and Insufficient

The US passed the Infrastructure Investment and Jobs Act in November 2021, which included $65 billion for broadband, the largest single federal investment in broadband infrastructure in US history. The BEAD program within it targets unserved and underserved areas with direct funding for fiber deployment.

This is real money and will connect real people who currently have no service. It is also about 25 years late. Countries that made these investments in the 2000s built the infrastructure that now carries 25 Gbit consumer connections. The US is still arguing about how to get 100 Mbps to rural areas.

The BEAD program also came with a requirement that funded networks offer at least one low-cost option, which is a partial acknowledgment that the market-only approach does not serve everyone. Whether the program will be implemented effectively depends on state-level decisions that are still being made, and some states have already shown more interest in protecting incumbent carriers than in maximizing coverage.

What 25 Gbit Means in Practice

A 25 Gbit symmetric connection is not a product most households need today. A single 8K video stream uses around 80-100 Mbps. A household with five simultaneous 4K streams and several active work-from-home connections might use 500 Mbps on a busy day. 25 Gbit is infrastructure headroom, not current demand.

But that is exactly the point. Infrastructure is built ahead of demand, not to satisfy current usage. You do not plan a highway for today’s traffic. Switzerland is building for 2035 by deploying 25 Gbit infrastructure in 2024. The US is still arguing about whether to fund gigabit connections to areas that currently have 10 Mbps DSL.

The underlying technology on init7’s Fiber7-X2 product is IEEE 802.3cm, 25 Gbit Ethernet over multimode fiber, running over the passive optical network infrastructure that Swisscom and municipal utilities have been deploying for years. The standard exists, the equipment exists, the fiber exists. The missing ingredient in most of the US is not technology. It is the political will to treat the physical layer of telecommunications as infrastructure rather than as a product to be monetized by whoever got there first.

Switzerland built the pipe. init7 sells the water pressure. The distinction matters enormously, and the US has spent 30 years pretending it does not.