IP Pool Depth for Social Accounts: How Many IPs You Really Need

If you manage social accounts, you have probably been told to look for the biggest IP pool you can find. That advice comes from the scraping world, and for accounts it is mostly backwards. An aged Instagram, TikTok or Facebook profile does not want a new address every hour; it wants the same carrier address in the same city for as long as the carrier will keep it there. New accounts being warmed up are a different story, and that is where the depth of the pool behind your line starts to matter. This guide walks through what a mobile IP pool really is, how many addresses an account manager actually needs, and how to read pool depth through the only lens that counts for this work: keeping accounts safe.

What is behind the address your account sees

Each ProxySocial line is one modem with one SIM in a rack we run in one of eight US cities, on AT&T, T-Mobile or Verizon. When the SIM connects, the carrier's gateway for that region gives the session a public address from the block that gateway manages. That block is the pool for your line. It is finite, it belongs to the carrier, and it is tied to the city the gateway serves, so a Verizon line in Miami and a Verizon line in Chicago draw from different sets.

The address is shared with ordinary phones through carrier grade NAT. To Instagram or TikTok, traffic from your line looks like traffic from a handset in that city on that carrier, which is exactly why mobile addresses are so kind to accounts. What the platform sees is not a private address that belongs to you, and it never was.

Aged accounts want a home, not a pool

An established account has a history: a device, a city, a carrier, a rhythm of logins. The safest thing you can do for it is keep all of that steady. Put it on a line in the city it has always lived in, on the stickier carrier available there, and leave the address alone. Do not schedule rotations. On a lightly loaded gateway the same address will stay with your line for days or even weeks, and if the line briefly drops and re-attaches the gateway will usually hand the same address straight back, because nobody else claimed it in the gap.

When customers write to us worried that an aged account's proxy has not changed address in a week, our answer is that this is the line doing its job. For this kind of account, an IP that never moves is a feature. The carrier's session timer, where one exists, is the only thing that will eventually change it, and a change once in a while looks exactly like a phone reconnecting after a night on wifi.

New accounts want fresh addresses, and that is where depth matters

Onboarding is the opposite job. When you create or warm up a batch of new profiles, you do not want three of them registering from the same address within an hour, and you do not want tomorrow's batch to land on addresses that yesterday's batch used. Here you rotate between accounts, and the question becomes whether each rotation hands you something the line has not held recently. That short-term freshness is one axis of pool depth. The other axis is long-run depth, how many distinct addresses the line reaches over a quarter, and the two do not always move together.

For onboarding, freshness within the day matters most. A line that yields a new address on most rotations lets you space account creations across the block naturally. As a rule of thumb, a single line rotated many times a day tends to turn up a few hundred distinct addresses in its first week and low thousands over a quarter, with the fastest discovery in the first days. The exact number depends on the carrier, the city and how often you rotate, so ask live chat which line in your metro is currently rotating freshest before you start a batch.

How many addresses does one account manager really need?

Fewer than you have been told. Count your accounts by type rather than adding up a wish list. Every aged account that matters wants its own sticky line, or at most a small number of accounts that plausibly belong to the same person sharing one line, in the same city, with rotation switched off. The onboarding pipeline wants one fresh-yielding line per metro you register accounts in, rotated between profiles. That is usually a small mix rather than a fleet. The pool depth question only applies to the onboarding lines, and there the honest measure is repeats within the day, not the grand total.

The trap is treating the headline pool as a safety margin. A vendor-wide count of millions is the sum of every gateway on every carrier the vendor has hardware on, and your line only touches one. What protects an account is that its address is a carrier address in a consistent city, held for a sensible length of time, with a device and behaviour that fit. A giant pool adds nothing to that.

• One aged account, or a handful that plausibly share a person: one sticky line in the account's city, no scheduled rotation.
• Account creation and warm-up: one fresh-yielding line per metro, rotated between profiles, checked for same-day repeats.
• Engagement or research tasks that never log in: any line, rotated as often as the task wants.
• Recovery after a checkpoint: keep the same line and address if you can, since changing everything at once looks worse.

Why rotating more does not make accounts safer

Rotation settings do not add addresses to the block. If a line is repeating, turning the timer from ten minutes to two minutes gives you more repeats, not fewer. Worse, for logged-in accounts, frequent address changes are themselves a signal, because a real phone does not hop between carrier addresses every few minutes. The lever that changes what a line yields is carrier and city. Broadly, across our racks, Verizon is the carrier most likely to give you something new on each rotation, AT&T is the one most likely to keep an address or bring an old one back, and T-Mobile depends heavily on which metro the line is in. The same carrier behaves differently in Houston than in Los Angeles.

If a rotation returns the address you just had and you actually wanted a new one, wait a little longer before reconnecting, or rotate again. Support can lengthen the detach time on your modem so the gateway is more likely to reassign the address while you are offline. And if a line's block is simply too shallow for the batch size you run, switching carrier in the same city or moving the line to another metro is free.

You are not hiding in a crowd

Sharing a gateway address with many phones is not anonymity, and it is not what keeps accounts safe. Platforms do not ban carrier addresses lightly because doing so locks out real subscribers, but they attribute activity by account, device fingerprint and behaviour. Two of your accounts on one browser profile, a burst of follows, or a login from a city the account has never seen will get noticed regardless of how many other people share the address.

What the mobile address gives you is trust, not cover. The carrier ASN reads as a real phone network, the geolocation reads as a consistent US city, and if you keep each identity on its own device profile with its own line and a human pace, that trust holds. Pool depth is an asset for onboarding and a non-issue for aged accounts. It is never a hiding place.

What to do next with ProxySocial

Sort your accounts into the two piles, aged and new, and buy lines for the piles, not for the headline. Put an aged account on one line in its own city, from five dollars a day, and watch it for a week with rotation off; the address staying put is the result you want. For onboarding, run one line for a day, rotating between profiles, and count how many rotations came back with an address from earlier that day. If that number is higher than your batch can tolerate, open live chat and ask which carrier in your city is rotating freshest this week. Switching carrier or moving the line to another of our eight metros is free, and support can adjust the detach time if you keep landing on the same address.

Frequently asked