What a Cheap VPN Is Economising On
A VPN service is not software you buy once; it is a standing commitment to move other people’s traffic and to hold addresses that other people’s traffic can come out of. Both of those cost money every month, in proportion to how much you use them, which means a very low price is not a discount on a fixed product — it is a statement about how much capacity and how many addresses each customer is being allocated. Understanding the cost structure tells you which corners are being cut before you find out the hard way.
No service is named, priced or recommended on this page. The point is the shape of the economics, which stays true whatever the current offers happen to be.
The costs that recur every month
Five of these matter, and they behave very differently from each other.
Transit and bandwidth. Every byte you send arrives at a machine in a data centre and leaves again toward its destination, and somebody is billed for both legs. This is the one cost that scales directly with how much you actually use the service, which is why heavy users are structurally unprofitable on a flat rate and why services care a great deal about how many of their customers are heavy.
Machines and their capacity. An exit is a real computer with a finite processor, and encryption is work. The number of customers who can be active on one machine before it degrades is a real number, and packing more of them in is the cheapest possible way to reduce cost per customer.
Addresses. Exit addresses in the older, universally supported address family are a scarce leased resource with an ongoing cost, and the supply is genuinely constrained rather than artificially so. A service needs enough of them to spread its customers across, and needs to keep replacing the ones that stop working well. This is the cost most invisible to buyers and the one that most shapes the experience — the tradeoffs between arrangements are in shared vs dedicated VPN IP addresses.
Abuse handling. Anything that comes out of a shared address is attributed to the address. Somebody has to answer complaints, keep ranges from being widely refused, and retire addresses that have become unusable. This is unglamorous staffed work with no marketing value, so it is the easiest thing to under-resource.
Engineering and support. Clients for several platforms, kept working as those platforms change, plus humans answering questions. Both are salaries, and salaries do not fall when you add customers.
Why the address line is the interesting one
Bandwidth is a commodity and machines are rentable, but a well-behaved address is neither. Its value depends on its history: what has been done from it, who has decided to refuse it, and whether the range it belongs to has a reputation. That history is a shared asset spoiled by whoever is careless with it.
A service running thin on addresses has two options, and both are visible to you. It can put more customers behind each one, which makes each address more likely to be flagged, or it can accept degraded addresses, which means sites treat you as suspicious for reasons that have nothing to do with you. Neither shows up on a comparison table, and both explain most of the “why does this site keep challenging me” experience. Which organisation actually holds the range you are exiting from is something you can look up yourself — see what an ASN and reverse DNS reveal about your connection.
What a very low price implies about funding
There are only a few honest answers to how a low price covers those costs, and it is worth being able to recognise which one you are looking at.
Most customers use very little. Averages carry a flat rate. This works and is not a criticism; it simply means the service is priced for light use and your experience depends on which kind of customer you are.
Money arrives up front for a long period. A long prepayment is cash now against costs spread over years, which is a legitimate financing choice and also a bet that you will not need to leave. The risk transferred to you is that the operator, the ownership, or the terms may change while you are locked in.
Capacity per customer is thin. The most common answer. Fewer addresses per customer, more customers per machine, and no slack when a location gets busy.
Something other than your subscription is paying. If a price is low enough that none of the above accounts for it, the remaining possibilities involve revenue you are not the source of, which is a different arrangement than it appears to be. The versions of that arrangement are laid out in what a free VPN is actually being paid with.
The tradeoffs you should expect, as categories
Not predictions about any product — just where thinning tends to show up, so you know what to watch.
- Crowded exits at busy times, felt as unsteadiness rather than a lower headline figure.
- Addresses that services distrust, producing challenges, refusals and endless verification.
- No option for an address of your own, which rules out some uses entirely.
- The second address family handled carelessly, either dropped or leaked rather than deliberately carried.
- A slower client, updated less often, which is where platform-change breakage lands.
- Support that is documentation only, fine until the problem is not in the documentation.
- Long commitments as the only way to reach the advertised rate, which converts a decision into an inertia.
Where economising is entirely reasonable
Cheap is not a synonym for bad, and the honest position is that most people are light users for whom thin capacity is invisible. If what you want is for the network in front of you to stop seeing your destinations, on one device, for an hour a day, none of the above will trouble you and paying more buys headroom you will never enter.
The mismatch appears when a low-cost service is asked to do a job that depends on the parts that were thinned — anything address-sensitive, anything continuous, anything where being treated as suspicious would matter.
Why there is no cheapest pick here
We have not subscribed to these services, so we have nothing to report about how any of them actually behaves under load, and we do not publish prices because a price stated on a page is stale almost immediately and reads as a recommendation regardless of how it is worded. A ranked list of budget options would be an invitation to buy on the strength of testing that was never done, which is the mechanism by which this genre stopped being useful.
What is durable is the cost structure. Learn it once and you can interrogate any offer yourself, which is the part a list of names never gives you. The rest of the decision sequence is in how to choose a VPN.