X Says the Follow Limit Is 5,000. Across 325,553 Accounts, the Real Wall Is 7,500.

Everyone who has hit it remembers the number: 5,000. That is the follow limit X has documented for years, and its help page still says so today — "Every X account can follow up to 5,000 accounts," after which "you may need to wait until your account has more followers," governed by a ratio the page declines to publish.

We had 325,553 real public X profiles in a search index, each with an exact following count. So we checked whether 5,000 is actually where people stop.

It isn't. It isn't even close.

5,000 is a number nothing happens at

If 5,000 were a hard cap, the following-count distribution would pile up against it — a spike at 5,000 and a cliff just after. Here is what the region actually looks like, in bins of 250:

4,000–4,2491,158

4,500–4,749945

5,000–5,249702documented cap

5,500–5,749520

6,000–6,249403

6,500–6,749324

7,000–7,249307

7,500–7,749257

8,000–8,249171

8,500–8,749163

The curve slides straight through 5,000 as if it weren't there. Exactly 4 accounts in the whole corpus follow exactly 5,000 — fewer than sit at 4,999 or 5,001. Whatever the help page describes, the number 5,000 leaves no fingerprint in real accounts at all.

The wall is at 7,500 — and you can prove it with the small accounts

To find the real cap, look at accounts that can't benefit from any follower-ratio exception: the small ones. If your follower count is low, a ratio rule can't lift your ceiling, so wherever these accounts stop is the flat floor everyone gets.

They stop at 7,500.

A hundred and fifty thousand accounts too small for any ratio bonus, and not one of them follows more than 7,508. The ceiling is flat, hard, and sits 50% above the documented 5,000. The distribution above shows people freely following 5,000, 6,000, 7,000 — right up until a wall that the help page doesn't mention.

There is even a faint pile-up exactly on it. Nineteen accounts in the corpus follow precisely 7,500 — nineteen times the density of neighbouring values, the only such spike anywhere in the entire following distribution.

We checked the wall was still standing

A spike in a months-old crawl could be a fluke of when we happened to look. So we pulled a sample of the accounts sitting at 7,500 live, straight from their profiles, and compared:

All twelve accounts we re-checked were still parked between 7,487 and 7,504, months later. That is what a live ceiling looks like: accounts drift a few follows above or below it as they add and drop people, but they cannot climb away from it. It holds for a former Premier of British Columbia with 86,875 followers and for an account with 236 followers alike — and for verified and unverified accounts equally, so it is not a paid-tier perk. It is the floor.

The "1.1x your followers" rule is folklore

Ask around and you'll be told the ratio rule is simple: you can follow up to about 1.1 times the number of people who follow you. It's repeated everywhere. It is also not something X has ever published — the help page has said the ratio "is not published" verbatim since 2012 — and the data breaks it:

• 9.4% of the 8,860 accounts following more than 5,000 people follow more than 1.1x their own follower count.
• @ZenvoAuto follows 7,500 people with 236 followers — 32 times the folklore limit.
• @drewdennison, CTO of a security company, follows 7,500 with 1,253 followers — 6 times it.

The honest version is narrower and stranger than the rule of thumb: there's a flat floor around 7,500 that anyone can reach regardless of follower count, and only above 7,500 does your follower count start to matter. The accounts following hundreds of thousands or millions — @yokoono follows 786,898; one account follows 1,395,626 — are the tail where a real ratio finally applies. Below 7,500, the "ratio" is irrelevant, because the wall gets to you first.

Almost nobody follows many people anyway

All of this is a story about the top ~2% of accounts, because the striking thing about the following count is how small it usually is. The median account in our data follows 314 people. And that number barely moves no matter how large the account gets:

under 1k followers128

1k–10k493

10k–100k568

100k–1M502

over 1M followers401

An account with five million followers follows, at the median, fewer people than one with fifty thousand. Gaining an audience does not make anyone follow more accounts — if anything, the largest accounts follow slightly fewer. The follow limit is a ceiling almost no one is anywhere near; the median user is using 4% of the documented one.

What this shows, and what it doesn't

What we think holds up: the documented 5,000 follow limit does not describe where accounts actually stop — there is no cliff there. The real flat ceiling is ~7,500, demonstrated by 151,189 small accounts none of which exceed 7,508, and confirmed live months after the crawl on every account we re-checked. The widely-repeated "1.1x followers" ratio rule is broken by roughly one in ten accounts above 5,000. And typical following counts are tiny and flat across the follower range.

What limits it:

• This is a seeded corpus, not a random sample of X — 325,553 accounts discovered from independent public sources (people notable enough for a Wikidata entity, public GitHub developers, founders, journalists) and then refreshed against their live profiles. It skews toward established and Western accounts. The ceiling we measure is a platform rule, so it should generalize; the typical following counts describe this population.
• We can't see X's internal rule, only its effect. Our reading — that the base cap is now ~7,500 rather than the documented 5,000 — is an inference from where accounts pile up and stop. X may compute it differently under the hood; what we can say for certain is that 5,000 is not where the wall is.
• Following counts are a single snapshot and drift over time. That is exactly why we re-checked the 7,500 accounts live; the numbers moved by a few follows, not by thousands.
• The ~7,500 figure has slop. Accounts sit at 7,487–7,508, not all at a clean 7,500, consistent with a limit you bump against and get pulled back below rather than a hard integer gate.

None of this is a scandal — a follow cap is mundane platform plumbing. It's just that the number in the documentation and the number in the data are not the same number, and 325,553 accounts are quite sure which one is real.

Sources

Methodology

We built a search index of 325,553 unique public X profiles, seeded from independent public sources and refreshed against each account's live profile through Crawlora's X profile endpoint. Each record holds follower count, following count, lifetime posts, join date and verification flag. The full corpus was pulled by paginating the X users dataset across account-creation date windows and reconciled against the dataset's facet counts — the census matched the independent aggregate to within 0.01%.

To locate the real cap we isolated accounts whose follower count is too low for any ratio-based exception to raise their limit (under 2,000 followers) and took the maximum following count in that group. To confirm the ceiling was current rather than an artifact of crawl timing, we re-fetched a sample of accounts sitting at 7,500 live and compared. Following counts are exact integers from the profile JSON, not the rounded values X's web profile displays.

Update, August 2026: a new discovery channel (Common Crawl) has since added 324,523 more profiles to this index with a very different composition.

We publish aggregates only — no individual profile records beyond the handful of public accounts named above to make the live re-check verifiable. Want to run your own cut? The how to scrape Twitter/X guide walks through the exact calls, the X users dataset is queryable directly, and pricing covers the free tier. For the companion study on who posts the most, see the most prolific poster on X is a machine.