What 'X monthly active users' actually counts

The number in any headline about X monthly active users is rarely the same metric twice. Twitter swapped MAU for monetizable daily active users in Q1 2019. X stopped filing with the SEC in October 2022. The SpaceX S-1 in May 2026 gave the first audited figure in over three years: 550 million MAU, 8.3% under Musk's stated 600 million.

X Monthly Active Users: One Label, Three Different Definitions

The phrase has pointed at three incompatible measurements in seven years. Before Q1 2019, monthly active users meant any logged-in account that touched Twitter within a 30-day window. From Q1 2019 through Q2 2022, the headline metric was monetizable daily active users: logged in, that day, on a surface able to serve ads. After October 2022 there was no published metric at all, only executive statements with no methodology attached.

The size of that definitional shift is visible in the quarters where both numbers existed. Twitter's last reported MAU was 330 million in Q1 2019. Concurrent mDAU was 152 million, less than half. Go back one quarter and the gap is wider: Q4 2018 MAU of 321 million against the first disclosed mDAU of 126 million, a 60% gap inside a single quarter of the same product.

Which is why the charts on stats aggregator sites are, in a strict sense, broken. A line running continuously from pre-2019 MAU data through post-2019 mDAU data plots two different questions on one axis. The apparent plateau in those charts is a measurement artifact, not platform behavior. Nothing about user activity changed on the day the definition did.

This is not academic if you buy ads. When you open X Ads Manager and size a B2B segment, US senior professionals in SaaS, the number the platform returns is drawn from ad-eligible accounts, not total MAU. In practice the addressable B2B audience on X is often 40 to 50% smaller than the headline user number suggests, because a large share of logged-in accounts sit on third-party clients, read-only sessions, or surfaces that never serve an ad.

So the first question to ask about any X user figure is not whether it is big. It is which of the three metrics produced it.

What mDAU Counts and What It Excludes

mDAU has a precise legal definition, filed with the SEC. From Twitter's Q2 2019 10-Q: 'Twitter users who logged in and accessed Twitter on any given day through Twitter.com or Twitter applications that are able to show ads.' Three conditions, all required. Logged in. That day. On an ad-capable surface. Miss one and you are not in the count.

The exclusion that mattered most was third-party auto-polling. Under the old MAU, an app that pinged Twitter's servers in the background on a schedule, with no human taking any action, could register that account as active. A phone in a drawer could count. mDAU removed the whole category, and with it a phantom audience that had been inflating every estimate built on top of MAU, B2B targeting estimates included.

The caveat in the same filing is the part aggregators never quote. Twitter wrote that mDAU 'is not based on any standardized industry methodology and is not necessarily comparable to similarly-titled measures presented by other companies.' The company that invented the metric stated, in a government filing, that you cannot use it to compare X against anyone else. People do it constantly anyway.

We build automation tooling for these platforms, so here is a wrinkle we have a direct view of. A session that logs in through a real browser on a home IP, loads the timeline, and can technically render an ad satisfies every mDAU condition, whether or not a human is watching in that moment. Browser-based automation of that kind contributes to mDAU. API-only tools do not. The spam estimate does not catch those sessions, because they pass every ad-eligibility test rather than failing one.

That cuts both ways for anyone reading the numbers. mDAU was built to answer an advertiser's question, not a researcher's. It counts ad-viewable sessions, and a session is not the same thing as a person paying attention.

Does X Count Bots in Its Monthly Active User Number?

Twitter's only official answer comes from the Q2 2019 10-Q: fewer than 5% of mDAU during the quarter were false or spam accounts. The same filing undercuts its own number in the next breath, stating that 'significant judgment was applied' and that 'the actual number of false or spam accounts could be higher than we have estimated.' No independent audit ever tested the figure.

In 2022, Musk put the number at a minimum of 20% of accounts while contesting the $44 billion acquisition, and demanded Twitter prove otherwise. SparkToro, measuring with its own methodology, landed on 19.42% fake or spam. The two figures agree closely with each other and disagree wildly with Twitter. That looks damning until you check the denominators.

Twitter's fewer-than-5% applied to mDAU: logged-in, ad-viewable, daily users. The 20% and 19.42% figures applied to all accounts, which includes dormant accounts, read-only accounts, accounts that only ever touched the API, and accounts that have not loaded an ad in years. Spam concentrates in the exact population mDAU excludes by construction. Both sides were measuring honestly and neither number refutes the other, which is why the dispute never resolved on the numbers.

The scale of the underlying problem is easier to read in enforcement volume than in percentages. In 2024 alone, X suspended 800 million accounts for spam. That is more than 40% above the platform's own total reported MAU. Account creation at that volume means the share of synthetic activity is a moving target recomputed continuously, not a stable percentage anyone can quote for a given year.

The honest answer to 'how many bots does X have' is that nobody outside the company knows. The last people with audited access to find out stopped filing in 2022.

Why X Stopped Publishing Quarterly User Data After 2022

The disclosure stopped for a mundane legal reason. Musk closed the $44 billion acquisition in October 2022 and took the company private, and private companies have no SEC obligation to publish audited user figures. Quarterly filings ended immediately. The mechanism that had forced a consistent, auditor-reviewed number out of the company every quarter simply switched off.

The last publicly disclosed mDAU was 237.8 million, for Q2 2022. Every X user figure that circulated between October 2022 and May 2026 came from a conference stage, an interview, or a post, not from a filing. None carried a methodology statement. Many were reproduced by press and aggregators in the same visual style as the audited figures that preceded them, which is how a self-reported claim acquires the look of a verified one.

Internally, the metric changed too. Musk replaced mDAU as the primary internal metric with 'unregretted user-minutes', later reframed as 'unregretted user-seconds'. The shift goes from counting heads to measuring session quality and time on platform. That metric has never been publicly defined or disclosed as a number, so there is no way to track it from outside the company.

The KPI change has a consequence you can observe in your own analytics. The formats that generate extended read time are threads with a reason to keep scrolling, quote posts with substantive commentary, and embedded media that auto-plays. Each draws distribution out of proportion to its raw engagement count. SocialNexis users see this in engagement analytics: a single high-dwell-time thread can outperform five quick-fire posts on impression volume. X's ranking weights time in view, the same thing the internal metric measures.

If you plan content against X, optimize for dwell time rather than click-through. The platform is not scoring you on the metric it used to publish.

The SpaceX S-1 Filed the First Audited X Monthly Active Users Figure in Three Years

The SpaceX S-1 IPO filing, dated May 20, 2026, contains the first SEC-audited X user figure since the October 2022 acquisition: 550 million monthly active users as of March 2026, with 520 million as of December 2025. Three months, 30 million users of growth, both numbers inside a document that carries securities-law exposure rather than a stage microphone.

Musk had been stating 600 million MAU for X since mid-2024. The filed figure is 8.3% below that, roughly 50 million users of difference that stayed invisible from outside until an unrelated IPO filing exposed it. That gap is the most useful calibration number in this guide: it tells you how much to discount the next unaudited figure you are handed.

Note which metric the filing used. It reverted to MAU, not mDAU. That makes the 550 million figure comparable to the 330 million from Q1 2019 and incomparable to the 237.8 million mDAU from Q2 2022. X has now run three different primary user metrics in seven years, and the 2026 number happens to belong to the same family as the 2019 one, which is the only reason any trend line is possible at all.

Before the filing, the single leadership-attributed daily figure was Linda Yaccarino's 225 million DAU, given at Vox Media's Code 2023 conference, which X later revised upward to 245 million. Neither version arrived with a definition. A DAU number that does not say whether it means ad-eligible sessions or any logged-in session is not something you can do arithmetic with, and an upward revision with no stated reason is not a correction anyone can audit.

Third-Party Estimates of X Daily Active Users Diverge by a Factor of Two

Two widely cited daily-active figures for X differ by nearly a factor of two, and the difference is entirely methodological. SimilarWeb's mobile app tracking showed 132 million DAU in June 2025. An all-platform estimate attributed to Financial Times reporting puts Q2 2024 at approximately 251 million DAU, a figure that reaches most readers secondhand through stats aggregators rather than through the original article. Neither is wrong. They count different surfaces.

SimilarWeb's number excludes desktop and web browser access, which makes it a floor rather than an estimate. That exclusion lands hardest on the exact audience B2B marketers want. Professional use of X skews toward desktop and web: longer reading sessions, a work machine, a browser tab open beside everything else. Mobile-only measurement undercuts that segment systematically. When you see an 'X is dying' chart built on app data, the first question is whether web sessions are in it.

A third figure gets cited alongside those two and measures something else again. DataReportal put X's potential advertising reach, the number visible inside X Ads Manager, at 586 million in January 2025, falling to roughly 557 million by October 2025. Subtract one from the other and the drop is about 29 million reachable ad accounts in nine months. The October endpoint is reported as an approximation, so treat the gap as roughly 30 million. It is still the closest thing to a near-real-time audience signal available outside the company.

So the three numbers in circulation, 132 million, 251 million and 557 million, are mobile app DAU, an all-platform DAU estimate, and ad-eligible accounts. Using any of them as 'X monthly active users' produces a figure nobody can verify or compare against last year's. The spread is not uncertainty about one quantity. It is three quantities wearing one label.

Use X Ads Manager Reach, Not Published MAU Numbers, for B2B Planning

For B2B planning, stop using published MAU. Since quarterly filings stopped in October 2022, the only audience number that updates in near-real time is the estimated reach field inside X Ads Manager, and it approximates the mDAU concept directly: logged in, ad-viewable, active. It is also the only figure scoped to your targeting rather than to the whole platform.

That scoping is the point. Platform-total MAU answers a question no media buyer has. The Ads Manager figure answers the one they do have: how many accounts matching this seniority, this industry and this geography can my ad reach today. A 550 million headline tells you nothing about whether fintech decision-makers in EMEA grew or contracted last quarter.

Because no segment-level data has been disclosed since 2022, the blind spot is specific rather than general. You cannot verify from any official source whether your target audience has expanded or shrunk. The practitioners who handle this well do something plain: save the targeting configuration for each key segment, record the estimated audience size, check it monthly. A declining estimated reach on your own segment is a more direct signal than any aggregator headline.

Pair that with a reach reality check, because audience size and audience attention are separate budgets. Organic reach for B2B accounts with 10k+ followers ran at 2.3% in 2025, down from 8.7% in 2020. Sizing a reachable segment is the first step. The share of it you touch without paying has fallen sharply over those five years, and no MAU figure has ever captured that.

None of this needs a data vendor. It needs the discipline to log a number monthly that the platform will not log for you.

Reading X Audience Data Without Getting Misled

Before citing any X user figure, identify which of four things it is: pre-2019 MAU, 2019 to 2022 mDAU, a post-2022 self-reported claim, or the SpaceX S-1 audited MAU. They are not interchangeable. A trend line assembled from more than one category describes a change in accounting, not a change in the platform.

For any third-party DAU estimate, check whether desktop and web sessions are included before treating it as a range. The 132 million mobile-only figure and the 251 million all-platform estimate are not the low and high ends of one estimate, even though both carry the label daily active users. Quoting the lower one in a B2B deck understates the exact audience the deck is about.

For audience sizing, use Ads Manager reach scoped to your parameters rather than platform-total MAU. The addressable B2B segment typically lands at 40 to 50% of the total MAU headline once seniority, industry and geography filters apply, and that is before accounting for how many of those accounts your organic posts will touch.

Treat every unaudited executive statement as a self-reported estimate with a known bias direction. The SpaceX S-1 established the size of that bias: 8.3% between the stated 600 million and the filed 550 million. When the next unaudited number appears, discount it, and say in your deck that you are discounting it.

The useful posture is not cynicism about X's size. It is precision about which question each number answers. The platform has changed its primary user metric three times in seven years and spent more than three of those years publishing none at all. Anyone quoting one confident figure across that whole stretch is quoting a definition change as if it were growth.

Frequently asked questions

What is the difference between MAU and mDAU on X (Twitter)?

MAU (monthly active users) counted any logged-in account that accessed Twitter within 30 days, including apps that contacted servers in the background without human action. mDAU (monetizable daily active users), introduced in Q1 2019, counts only users who log in daily and access a surface that can show ads. When Twitter switched, the MAU figure was 330 million while concurrent mDAU was 152 million, less than half, because mDAU filtered out non-ad-eligible access patterns.

When and why did Twitter stop reporting monthly active users?

Twitter announced on February 8, 2019 that it would stop reporting MAU starting with Q1 2019 earnings. The stated reason was to align external stakeholders around one metric that reflects delivering value to people every day and monetizing that usage. The switch also replaced a 330-million headline number with a 152-million one, because mDAU filtered out non-monetizable access patterns that had inflated the prior figure.

Does X count bots in its monthly active user number?

According to Twitter's Q2 2019 10-Q, fewer than 5% of mDAU were false or spam accounts, though the filing acknowledged 'significant judgment was applied' and 'the actual number could be higher.' In 2022, Musk estimated at least 20% of accounts were spam, but that figure applied to all accounts, not just mDAU-eligible ones. The two estimates use different denominators, which explains why they appear so far apart.

Why did X stop disclosing quarterly user metrics after Elon Musk's acquisition?

When Musk completed the $44 billion acquisition in October 2022, Twitter became a private company with no SEC obligation to publish audited user figures. The last publicly disclosed mDAU from any official filing was 237.8 million (Q2 2022). Since then, every user figure has come from executive statements at conferences or in media, not from audited filings. The SpaceX S-1 in May 2026 was the first exception to that pattern.

How do analysts estimate X daily active users without official figures?

Analysts use three main proxies: app analytics firms like SimilarWeb track mobile app sessions (132 million DAU in June 2025, excluding desktop); financial press sources like the Financial Times cite all-platform estimates around 251 million DAU; and X Ads Manager shows ad-reachable audience figures by targeting segment. All three differ significantly because they measure different access patterns, not a single consistent underlying metric.

What does 'monetizable' mean in X's mDAU definition, and what gets excluded?

Monetizable means the user accessed X through a surface that can show advertisements, specifically Twitter.com or official Twitter apps. The explicit exclusion from Twitter's Q2 2019 10-Q is third-party auto-polling: apps that contact Twitter's servers in the background without a human taking any action. Those background server pings had counted as 'active' in prior MAU calculations. Users on API access tiers that do not serve ads are also excluded.

How accurate is Elon Musk's claim that X has 600 million monthly active users?

The SpaceX S-1 filing (May 2026) provided the first externally audited X user figure since October 2022: 550 million MAU as of March 2026, 8.3% below Musk's stated 600 million. The same filing shows 520 million MAU as of December 2025. The gap of approximately 50 million users was invisible without external auditing. Musk's 600 million figure appears to be an upward-rounded self-reported estimate rather than an audited one.

What is X's 'unregretted user minutes' metric and how does it differ from mDAU?

After the 2022 acquisition, Musk replaced mDAU with 'unregretted user-minutes' (later called 'unregretted user-seconds'), measuring session quality and time spent rather than daily head count. Unlike mDAU, this metric has never been defined in a public filing or disclosed numerically. Its observable effect on the content algorithm is that posts with high dwell time, such as threads and embedded media, receive disproportionate distribution relative to their raw engagement counts.

How many fake or bot accounts does X have in 2025?

No verified current figure exists. Twitter's last official estimate was fewer than 5% of mDAU (Q2 2019 10-Q), a figure that was never independently audited. In 2024 alone, X suspended 800 million accounts for spam, a number exceeding total reported platform MAU by more than 40%. The suspension volume indicates the real share of fake activity is a continuously moving target, not a fixed percentage of any published user count.

How does X's audience size figure affect B2B ad targeting and reach planning?

For B2B advertisers, the figure that matters is the estimated reach inside X Ads Manager, not the headline MAU number. The ad-reachable audience dropped from 586 million in January 2025 to approximately 557 million by October 2025. When you apply B2B filters such as seniority, industry, and location, the resulting segment is typically 40-50% smaller than the total MAU headline, because many accounts are not ad-eligible or do not match professional targeting parameters.

Sources and further reading

• Twitter 10-Q Q2 2019: the official mDAU definition and spam account methodology
• Social Media Today coverage of the SpaceX S-1 X user data disclosure (2026)
• Twitter 10-K FY2021: the last full annual SEC filing before privatization