The numbers X stopped publishing, and what the outside data says instead

The last tweet volume the platform officially confirmed was 500 million a day, published in August 2013. Thirteen years later that figure still circulates as though it were current, which tells you most of what you need to know about the state of public data on X.

The company publishes far less than it once did. What fills the gap is third-party measurement, and those estimates disagree with each other in ways worth understanding before quoting any of them. The compilation of twitter statistics maintained by TwitterAPIs is one of the few that attaches a named source and a date to every line, which is the minimum bar for a figure to be quotable at all.

What the outside estimates actually say

X's own ad-planning tools showed an addressable audience of 586 million users worldwide in January 2025, according to DataReportal. Third-party analysis put the platform at an estimated 561 million monthly active users by July 2025, down from 586 million a year earlier, per Backlinko.

Daily usage tells a sharper story. About 132 million people used X daily across the iOS and Android apps in June 2025, a 15.2 percent year-over-year decline. Potential advertising reach fell by 33 million users, roughly 5.3 percent, in the twelve months to January 2025.

For scale, that audience equals about 7.1 percent of the global population and 9.9 percent of all adults aged 18 and over. The United States remains the largest market at roughly 104 million active users, ahead of Japan at 70.9 million. Advertising revenue was an estimated $1.94 billion in 2024.

The bot question, where the estimates diverge most

How much of X is automated is the most contested number on the platform, and the range is wide enough that the answer depends entirely on who is counting.

Twitter told regulators that fewer than 5 percent of its monetizable accounts, about 20 million, were fake or spam. A 2017 academic study estimated that up to 15 percent of active accounts were automated. A 2020 estimate placed bots at roughly 15 percent of all accounts, around 48 million. Analytics firm Sysomos estimated that bots created about 24 percent of all tweets.

Those are not contradictory so much as measuring different things. A share of accounts and a share of tweets are different denominators, and automated accounts post more than human ones by design, which is exactly why the tweet share runs higher than the account share. Anyone citing a single bot percentage without saying which denominator it uses is quoting a number that cannot be checked.

Why this is now a measurement problem

The practical consequence is that anyone who needs current platform data has to collect it rather than look it up.

That is a change in kind, not degree. Ten years ago a researcher could cite an official figure. Today the honest options are to use a dated official number and say so, to cite a third-party estimate and name the estimator, or to measure the slice you actually care about directly.

The third option is more accessible than it sounds. Targeted collection through twitter search operators narrows a query to a date range, an account set, an engagement floor or a language before anything is fetched, which turns an unbounded question into a countable one. Running that at volume is what a twitter scraper is for.

How to quote any of this responsibly

Three habits keep a citation defensible. Name the source and the date beside every figure, because "500 million tweets a day" is true of 2013 and unknown today. State the denominator on any percentage. And prefer a measured slice of the thing you care about over a platform-wide estimate you cannot verify, since a number you collected is one you can defend.

Official data regarding activity on X has been notoriously scarce since 2013, when the platform last confirmed a daily output of 500 million tweets. In the absence of primary updates, third-party estimates provide the clearest window into the platform's trajectory, though they depict a contracting audience. By mid-2025, estimated monthly active users sat around 561 million—a year-over-year decline—while daily active app usage dropped over 15% to roughly 132 million users. Despite these contractions, X retains substantial global reach, particularly in key markets like the United States and Japan, even as advertising reach and revenues have noticeably softened.

The divergence among third-party figures becomes most pronounced when examining automated activity, where metrics depend entirely on the chosen frame of reference. While official disclosures historically pegged fake or spam accounts at under 5%, academic and independent estimates place automated accounts closer to 15%, with bot-generated content accounting for nearly a quarter of all tweets. These figures are not inherently contradictory; automated accounts post at significantly higher frequencies than human users by design. Consequently, citing a single bot statistic without specifying whether the denominator represents total accounts or total tweet volume yields a functionally unverifiable metric.

Because macro-level metrics are both fragmented and difficult to verify, reliable analysis now requires direct data collection rather than passive citation. Researchers and analysts increasingly rely on targeted search operators and scraping tools to isolate specific date ranges, user sets, and engagement thresholds, converting unbounded platform questions into countable datasets. To quote data responsibly in this landscape, one must consistently attach explicit dates and sources to every statistic, clarify underlying denominators, and prioritize defensible, direct measurements over ambiguous platform-wide estimates.Each of those carries its own source and date on the TwitterAPIs compilation, which is worth checking before any of them is repeated, because several widely circulated figures in this category are years out of date.The platform got quieter. The measurement did not have to.