"Millions of IPs" Is a Vanity Metric: How to Test Real Pool Depth

Every mobile proxy provider has a big number on its homepage. Millions of IPs, tens of millions, the whole country at your fingertips. We have never printed one, and this page explains why. The number describes the carrier's address space, not the line you are paying for. What you actually buy is one modem with one SIM, attached to one carrier gateway in one metro, and the addresses that gateway is willing to hand that SIM. That is the pool that matters to you, and it is a great deal smaller than millions. The good news is that it is easy to check for yourself. Below is the test we suggest running on any line, ours or anyone else's, before you commit to a month.

Where the millions number comes from

Add up every address block that AT&T, T-Mobile and Verizon have registered and you get a figure in the tens of millions. That is the honest origin of the headline. A provider with modems on all three carriers can, in a loose sense, claim to touch all of it. The problem is that no single line can. A SIM on a Verizon modem in Houston gets its public address from the pool behind the Verizon gateway that serves Houston. It will never receive a Chicago address, and it will never receive an AT&T address, no matter how many times you rotate.

So the question a buyer should ask is not how many addresses the provider's carriers own in total. It is how many distinct addresses one line will actually show over the period you care about, and how often it hands you an address you saw yesterday. Those two numbers are what pool depth means in practice, and no homepage prints them because they change by carrier, by metro and by how the line is used.

The test: rotate, log, count

You do not need special tooling for this. You need a proxy, a way to trigger a rotation, an IP echo service and a spreadsheet or a text file. On our lines the rotation is a button in the dashboard, a rotation link you can call from a script, or an API call; underneath, a rotation is a data-session detach and re-attach on the modem and it finishes in seconds.

Run the loop for a full day if you can, and for at least a few hours. Rotate, wait for the modem to come back, fetch your public address, write it down with a timestamp, repeat. Pick one cadence and keep it, because a line rotated every two minutes will show a much larger count than the same line rotated hourly, and mixing cadences makes the numbers meaningless. When the day is over, count four things.

The first two give you yield: fresh addresses per rotation. The third tells you about short-term freshness, which is what verification and research work cares about. The fourth shows whether repeats are scattered or come in clusters, which usually points at a quiet moment on the pool rather than a shallow block.

• Total rotations you performed
• Distinct addresses seen, with duplicates removed
• Rotations that returned an address the line had already held that day
• Longest run of consecutive rotations without a repeat

What a good result looks like

On a healthy carrier line rotated many times a day you should expect a few hundred distinct addresses over a week and low thousands over a quarter. That is a rule of thumb, not a promise, and the exact figure depends on the carrier, the metro and your cadence. The shape of the curve is more telling than the total. In the first hours almost every rotation is new. By the end of the day the line has seen most of what its gateway routinely hands out and new addresses arrive more slowly. They never stop arriving entirely, because carrier pools shift as phones join and leave, but the pace settles.

A poor result looks different. If a line rotates fifty times and you count a dozen addresses, the gateway is handing that SIM a narrow block and hammering the rotation button will only churn it faster. If the repeats are back to back rather than spread out, the pool was simply quiet at that moment and the carrier handed the address straight back; wait a little longer between rotations and it usually clears. Support can also lengthen how long the modem stays detached before it re-attaches, which gives another phone the chance to take the address.

Compare lines, not providers

The most useful version of this test is running it on two lines at once. Same metro, different carriers, or same carrier in two metros. In our racks the carriers behave differently enough that this is worth a day of your time. Verizon lines tend to hand out a fresh address readily. AT&T lines tend to be stickier and revisit addresses more often, which is exactly what you want for a long-lived account and exactly what you do not want for a verification job. T-Mobile varies a great deal from one metro to the next.

Because of this, a provider claiming one depth figure for everything is telling you nothing. The lever you actually control is carrier times location. Rotation settings cannot create addresses that a gateway does not hold. If your test shows a shallow block, the fix is a different carrier or a different city, and on our side both moves are free.

Questions to ask any provider before you buy

Once you have run the test once you will know what to ask, and the answers separate providers that run hardware from providers that resell someone else's. Ask about a specific line, not about the company's carriers in general. Ask whether the modem and SIM are dedicated to you or shared with other customers, because a shared modem means your rotations are somebody else's rotations too. Ask what a rotation physically does and how long it takes.

Any provider who answers these with a pool size instead of a straight answer is selling you the headline.

• Which carrier and which city is this exact line on?
• Is the modem and SIM dedicated to me or shared?
• What does a rotation do on the hardware, and how long does it take?
• Can I switch carrier or city if the pool is shallow, and what does that cost?
• Can I keep a sticky session for as long as I want?

Do not mistake a shared address for a hiding place

One more thing the millions number invites you to believe: that a bigger pool means more anonymity. It does not. Every mobile address sits behind carrier grade NAT and is shared with many ordinary phones at the same moment. That sharing is why platforms are reluctant to ban carrier addresses outright, but it is not a crowd you disappear into. Platforms attribute by account, by device and by behaviour. The trust a mobile line brings comes from the carrier ASN and from consistent geography for each account, not from the number of neighbours on the address. Pool depth buys you fresh addresses when you need them; it does not buy cover for sloppy account handling.

Run the test on a BetterProxies line

Our plans start at a single day, so the test above costs you a day and an afternoon of attention. Run it against the real target you care about rather than against a checker, because a checker cannot tell you whether the line works for your job. If the count looks thin, open the live chat and tell us the carrier, the city and what you saw; we can say which line in your city is rotating freshest right now, switch you to another carrier, or move the proxy to any of our eight metros, from New York to Phoenix, at no charge. We would rather you count for yourself than trust a number on a homepage, including ours.

Frequently asked