Why a VPN's Costs Rise With Every User

A VPN has the cost structure of a utility, not of software. Writing an app is a one-time cost that spreads thinner the more people install it; carrying their traffic is a recurring cost that grows with every user and every gigabyte. That is why free VPN tiers have limits and why “free and unlimited” is a claim worth examining rather than a feature.

Understanding this is the whole basis of the site: once you know that the cost is real, continuous, and proportional to usage, the only interesting question about any free service is where the money comes from.

The costs that behave like software

Some of a VPN provider’s spending really does spread out as it grows. Building the client apps, maintaining them across platforms, designing the protocol integration, the website, the brand — all of that is roughly the same amount of work whether ten thousand or ten million people use it. This is the part of the business that looks like normal software economics, and it is the part that makes a large user base attractive.

If a VPN were only apps, a free tier would be nearly costless and there would be nothing to explain. But the app is the thin part of the product.

The costs that behave like a utility

Bandwidth. Every byte a user sends goes into the tunnel and out again, so the provider pays for it twice — once inbound, once outbound. Transit and egress are sold by volume or by sustained rate, and neither gets cheaper because more people are doing it. A user who streams video is meaningfully more expensive than one who checks email, and no amount of growth changes that relationship.

Server capacity. Encrypting and routing traffic consumes CPU and network interface capacity. More concurrent users means more machines, in more places, each with a monthly bill.

IP address space. Exit addresses are leased or purchased, and they are a genuinely scarce input rather than something a provider generates. This is its own line item, covered in why IP address space is a line item.

Operations. Someone has to be on call, someone has to handle abuse complaints, someone has to answer support tickets. Those scale with users too, just less smoothly — covered in the hidden cost of a free tier.

What a single new free user actually adds

Not much on its own. A few gigabytes a month, a fraction of a server’s capacity, a share of an IP address. The number is small.

The problem is that it is a positive number that recurs every month, and it does not fall as the user base grows. Multiply a small recurring cost by a large number of users and you get an operating expense that has to be funded by something. A free user is not a rounding error; a free user is a subscription the provider is paying on your behalf.

This is also why usage distribution matters more here than in most businesses. Any service where cost is driven by consumption will find that a minority of users account for a large share of the bill. Providers do not need to know which users those are in advance — they only need a cap, and the cap does the sorting. How that number gets chosen is in how a freemium VPN decides where to put the cap.

Where the real efficiencies come from

Providers are not helpless about this. There are several genuine levers, and none of them is “scale makes it free”:

Peering and transit negotiation. A provider moving a lot of traffic can get better rates and can peer directly in some locations, cutting out a paid hop. This is a real discount, but it is a discount on a bill that still grows.

Location arbitrage. Bandwidth and hosting cost very different amounts in different markets. This is one reason free tiers offer a short list of server locations: the list is chosen partly for cost, not only for coverage.

Oversubscription. This is the big one. Not everyone is online at once, so a server can be provisioned for peak concurrency rather than for the total number of accounts. Every network in the world is built this way, including your home broadband.

Prioritisation. Once you have oversubscribed, you need a rule for what happens when demand exceeds capacity. Deprioritising free users during congestion is the standard answer, and it is the mechanism behind the observation that free tiers slow down at busy times.

What this predicts about any free service

Three things follow directly, without needing to know anything about a specific provider:

1. A free tier will be limited in some way — volume, locations, devices, priority, or features. If it is not limited, the cost is being recovered by a mechanism you have not identified yet.
2. Limits are not stinginess. They are the shape the economics forces. A provider that removed them would be spending more on non-customers every month, indefinitely.
3. The funding model is discoverable. Something is paying that recurring bill: paying subscribers, advertisers, data buyers, proxy customers, donors, or a parent company. The six recognisable patterns are enumerated in how free VPNs make money.

Using this as a reader

You do not need cost figures to apply any of it, and you should be suspicious of anyone quoting per-user costs they cannot source. The useful version is directional:

• Heavier usage is more expensive to serve, so generous free allowances are a bigger commitment than generous feature lists.
• Unlimited free is the expensive promise, and therefore the one that most needs an explanation. See the arithmetic of an unlimited free VPN.
• A provider with paying customers has an obvious funder for its free tier. One without needs a different answer, and there usually is one.

None of this means free services are bad. It means free services are funded, and the funding is the thing worth knowing before you install anything. Where free is straightforwardly the right choice is covered in when a free VPN is actually fine.