Elon Musk completed his acquisition of Twitter for $44 billion in October 2022. ³ At that point the company was a public firm whose last regulatory filing—an SEC Form 10-Q for the second quarter of 2022—reported total revenue of about $1.18 billion, a net loss of $270 million, and 237.8 million monetizable daily active users. ⁴ Advertising then accounted for 91% of revenue. ¹

Over the following two years the company was renamed and restructured. Twitter was rebranded to X on July 23, 2023, and its web domain moved from twitter.com to x.com on May 17, 2024. ¹ Because the company went private, no audited post-acquisition financial statements are available in the record; the figures that follow come from a mix of one SEC filing predating the deal, business-press estimates, third-party analytics firms, and investor valuations. This report distinguishes throughout between what is directly documented and what is estimated or forecast.

Two independent origins converge on the headline revenue trajectory. Wikipedia, aggregating secondary sources, reports that revenue fell from a $5 billion peak in 2021 to $2.5 billion in 2024. ² Investopedia independently reports 2024 revenue of $2.5 billion, a 13.7% decrease from 2023. ³ CompaniesMarketCap, which for the pre-acquisition period draws on Twitter's public filings, puts 2021 revenue at $5.07 billion, agreeing with Wikipedia's peak figure within 1.4%. ¹ The analyst classifies both the 2024 total ($2.5 billion) and the 2021 peak (~$5 billion) as convergent across independent sources, giving those two anchor points high confidence.

The interior of the decline is documented mainly through advertising. Fortune reports global ad revenue fell from $4.5 billion in 2022 to $2.2 billion in 2023 and to approximately $2 billion in 2024. ¹ Wikipedia states the advertising sector fell by $2.8 billion over the period. ² The analyst notes an internal-consistency wrinkle: a $2.8 billion ad decline exceeds the $2.5 billion total revenue decline from peak, which implies non-advertising revenue must have risen by roughly $0.3 billion to net out—arithmetically consistent, but only if offsetting non-ad growth occurred. ² A second discrepancy is that Fortune characterizes the 2022-to-2023 ad drop as 46%, while the underlying figures ($4.5B to $2.2B) compute to 51.1%; the stated percentage and the stated dollar figures do not fully agree. ¹

The United Kingdom subsidiary offers the one place where year-by-year filings exist, and it fell faster than the global figure. Fortune reports UK operations revenue fell 66% in 2023 (from $282.9 million in 2022 to $95.2 million) and a further 58% in 2024 (to $39.8 million). ² The analyst verified both percentages against the stated dollar amounts (66.35% and 58.19%, respectively). ² These are the sharpest declines in the record, but the analyst cautions that UK revenue reflects one regional subsidiary and cannot be extrapolated to global performance without weighting by market size. ²

Against this decline sits one contrary data point. Investopedia reports X generated $2 billion in revenue over the first nine months of 2025, with quarterly revenue 17% higher than the year before. ¹ This is a single-source, secondary figure with no primary documentation, and it is a nine-month partial-year number that cannot be directly compared to the full-year 2024 total of $2.5 billion without annualization. ³ Any growth it represents is measured from a base roughly half the 2021 peak.

Open: What did audited X Corp full-year 2023, 2024, and 2025 financial statements report, broken out by advertising versus subscription revenue?; Which figure—Investopedia's implied ~$1.7B or Fortune's ~$2B—correctly states 2024 advertising revenue, and does the gap reflect inclusion of subscription revenue?

The user-base picture is consistent in direction and fragmented in method. The only official baseline is the 237.8 million monetizable daily active users in Twitter's final pre-acquisition 10-Q. ⁴ No comparable official metric was disclosed afterward, so post-acquisition figures come from third-party analytics and surveys measuring different things. The analyst stresses these are non-comparable denominators: monetizable daily users, monthly mobile users, and potential advertising reach cannot be placed on the same trend line. ⁴

With that caveat, the estimates all point down. Sensor Tower measured a 16% decline in global active daily mobile users by September 2023, and SQ Magazine cites Sensor Tower again for 252 million monthly active mobile users averaged across Q4 2024 and 145 million mobile app downloads in 2024, an 11% year-over-year fall in downloads. ² Wikipedia's one-year-post-acquisition snapshot reports a 30% decline in active users, a 14% decline in website traffic, and a 38% decline in app downloads—though the analyst flags that these four figures within a single claim use different metrics without a stated shared baseline, making them an apples-to-oranges bundle. ¹

Survey data on US adults diverges by house. Pew Research Center found 21% of US adults used X in 2023-2024, down from 23% in 2021—a two-point drop over roughly three years. ¹ Edison Research found a steeper fall, from 27% to 19% between 2023 and 2024, which it characterized as a 30% usage drop. ¹ The analyst treats these as divergent: both agree on decline, but Edison's 8-point drop over one year and Pew's 2-point drop over three reflect different methodologies and timing rather than a settled magnitude. ² Regionally, Ofcom reported in November 2023 that X had lost 3 million monthly UK visitors, down from 26.8 million since the takeover. ¹ As of January 2025, Kepios put X's potential global advertising reach at about 586 million—a figure that may include logged-out audiences and is not a user-base count. ¹

Open: What do full time-series DAU/MAU figures from Sensor Tower or Similarweb show across all major geographies from 2022 through 2026?; Did X disclose any official user metric after Q2 2022, and if not, why has no regulator or auditor required one?

The organizational changes are the best-triangulated part of the record. Wikipedia reports the workforce fell from about 7,500 to approximately 1,500—an 80% reduction—through layoffs and resignations. ³ Aura independently reports a cut from around 8,000 to about 1,500 by April 2023, a nearly 80% reduction. ¹ The two starting counts differ by roughly 500, which the analyst attributes to the timing of the pre-acquisition measurement; both agree on the ~1,500 endpoint and the ~80% magnitude, and the arithmetic (6,000 of 7,500) confirms 80%. ³

Product and policy changes are documented descriptively rather than financially. In February 2023, X eliminated free API access, shutting down third-party clients including Tweetbot and Echofon. ² The following month it introduced paid API tiers at $100, $2,500, and $42,000—which Elfsight characterizes as roughly a 100x jump in entry price. ² These moves closed off free ecosystem access and created a new, if narrow, revenue channel, but the record contains no figures on how much API revenue they produced. The analyst rates SQ3 coverage weak-to-moderate: the headcount cut is well-corroborated, but no source quantifies the financial impact of specific policy changes. ⁴

Open: What revenue did the paid API tiers generate, and how many commercial customers adopted each tier?; How did the workforce reduction map to specific functions—engineering, trust and safety, sales—and did headcount change again after April 2023?

Advertising was the engine of the old business and the epicenter of the decline. It was 91% of revenue in Q2 2022 but approximately 68% by 2024, an Investopedia estimate from which the ~$1.7 billion ad figure is derived rather than directly reported. ³ The analyst cautions this shift reflects both an absolute ad collapse and a shrinking total denominator, so the falling percentage should not be read as successful diversification into other revenue. ³ On dollars, EMARKETER reports X ad revenue plummeted 51.7% in 2023, consistent in direction with Fortune's global ad figures. ² Wikipedia's one-year snapshot cites a 60% decline in advertising. ¹

One documented mechanism of advertiser flight appears in the record: on November 20, 2023, X sued Media Matters, alleging defamation over the watchdog's report on ads appearing next to antisemitic content. ¹ The lawsuit is a fact; the record does not quantify how many advertisers paused spending as a result.

The recovery case for advertising rests on a single forecast. EMARKETER predicts X's worldwide ad revenues will rise 16.5% year-over-year in 2025, which it frames as the first growth post-acquisition. ³ This is a prediction, not a measured outcome, and it is single-source for the 2025 projection. ³ It also follows directly from the 51.7% plunge of 2023, so a 16.5% gain would recover only a fraction of what was lost. ¹ There is a divergence on the 2024 ad-revenue level itself: Investopedia's ~$1.7 billion (68% of $2.5 billion) versus Fortune's ~$2 billion—a $300 million, roughly 15%, gap the analyst attributes to possible rounding, different fiscal endpoints, or Fortune's inclusion of non-advertising revenue. ³

Open: Which major advertisers paused spending after the Media Matters report, and how many resumed by 2025?; Did EMARKETER's forecast 16.5% growth materialize in reported 2025 results, and against what quarterly baseline?

X Premium was the flagship diversification effort, and the record shows it remains marginal. Investopedia reports X Premium made $11 million on mobile in its first three months after the November 4, 2022 relaunch. ² Searchlab.nl estimates that by 2026, X Premium had about 1.3 million paying subscribers worldwide—roughly 0.21% of the user base—generating $100-120 million a year, which it says covers less than 4% of total revenue. ¹ Both the subscriber count and the revenue figure are single-source dependencies on Searchlab with no independent confirmation in the record. ¹

The analyst's arithmetic surfaces a discrepancy in the framing: $100-120 million against $2.5 billion total revenue computes to 4.0-4.8%, so the claim of "less than 4%" understates the range slightly. ⁴ Either way, the contribution is small. For comparison, Twitter's own pre-acquisition subscription-and-other revenue was already $100.7 million in Q2 2022 alone—9% of quarterly revenue at the time. ¹ That baseline makes clear that a subscription line generating roughly $100-120 million annually has not materially reduced the platform's continued dependence on advertising, which still stood at about 68% of revenue in 2024. ²

Open: What is X Premium's total (not mobile-only) revenue, and how have subscriber counts and average revenue per subscriber changed year to year?; Has any independent source corroborated Searchlab's 1.3 million subscriber estimate?

The central question—what happened to revenue and users—splits into four testable explanations drawn from the analyst's hypotheses, each rated plausible on current evidence.

H1, revenue collapse: X's total revenue fell sharply from ~$5 billion to ~$2.5 billion by 2024, driven by advertiser exodus and not offset by subscriptions. If this were true, we would expect independent sources to converge on both the peak and the trough, ad revenue to fall in absolute terms, and subscription revenue to remain small. All three hold: the peak (~$5B, C-033 and C-009) and 2024 trough ($2.5B, C-002 and C-009) are convergent across independent origins, ad revenue fell from $4.5 billion to ~$2 billion, and subscriptions contribute at most ~5%. ⁴ No claim in the record contradicts H1. Its discriminating evidence—audited X Corp statements with a year-by-year ad-versus-subscription split—remains non-public, so the shape is well-supported but the interior detail rests partly on estimates.

H2, user-base decline: X lost users across metrics and geographies. If true, we would expect multiple independent analytics firms and surveys to show declines even if magnitudes differ. They do: Sensor Tower's 16% mobile daily-user drop, Ofcom's 3 million lost UK visitors, and Pew and Edison both showing US declines all point down, though the analyst's normalization warns the metrics are non-comparable in kind and the survey houses diverge in magnitude. ⁴ Direction is robust; precise scale is not.

H3, advertiser recovery in 2025: the worst of the exodus ended by late 2024 and ad revenue began recovering. If true, we would expect the 2025 forecast to be one of several converging indicators. Here the evidence is thinnest: the case rests essentially on EMARKETER's single-source forecast of 16.5% growth, supported circumstantially by the 17% quarterly revenue rise Investopedia reports and the near-par debt sale. ⁴ The debt sale is telling for creditor confidence but is backward-looking on solvency, not forward-looking on advertising; and all three indicators are UNGRADED and either forecast or single-source. H3 is the least-supported hypothesis, and its discriminating evidence—actual 2025 quarterly ad-spend data—is absent from the record.

H4, subscription irrelevance: alternative revenue streams contributed negligibly. If true, we would expect subscription revenue to be a low single-digit share against billions in advertising. The evidence supports this directly: ~$100-120 million against $1.7-2.0 billion in ads, with advertising still 68% of the total in 2024. ³ H4 is well-supported, subject to the single-source dependency on Searchlab. ¹

The hypotheses are not mutually exclusive. H1, H2, and H4 are mutually reinforcing and jointly well-supported; H3 is the contested frontier where the recovery narrative lives, and it is precisely there that the evidence is weakest.

Open: What discriminating evidence—audited segment financials or full-series third-party analytics—would confirm or reverse the tentative 2025 stabilization signal?

The evidence forces a clear conclusion on the core question. X is substantially smaller and less valuable than the company Musk bought. Two independent origins fix revenue at roughly half its 2021 peak by 2024; advertising, the dominant business, fell in absolute dollars; every measured user metric declined in direction even where magnitudes are contested; the workforce was cut about 80%; and Fidelity marked the equity down 72-79% from the purchase price. ⁴ These are the strongest-supported findings and the wording here is correspondingly firm.

The valuation figures deserve a note on their own limits. The Fidelity marks do not describe a straight-line collapse: the analyst's arithmetic shows a $19.8 billion implied value in October 2023 (-55%, C-006), a lower $9.24 billion implied in October 2024 (-79%, C-012), and a higher $12.5 billion in Q3 2025 (-72%, C-029)—a non-monotonic path consistent with quarterly mark-to-market of an illiquid stake rather than a clean trajectory. ⁴ The October 2024 and Q3 2025 marks are each single-source. ³

On the recovery question the evidence does not force a conclusion, and this report declines to reach one. The three pillars of the turnaround case—the 17% quarterly revenue rise, the debt sold at 98 cents on the dollar, and the 16.5% ad-growth forecast—are respectively a single-source partial-year figure, a backward-looking solvency signal, and an unrealized prediction. ⁴ Each may prove accurate; none is corroborated in the record, and all measure improvement from a depressed base. It is worth noting where documented interests bear on framing: X Corp leadership has an interest in demonstrating value recovery to justify the acquisition price, and the debt sale and forecast are the metrics that narrative leans on. ⁴ Symmetrically, Fidelity's valuations reflect its own required mark-to-market of an investment position. ³

(SPECULATIVE) One reading that the evidence permits but does not establish: the 2025 signals could mark either genuine stabilization from a smaller, leaner cost base—the 80% headcount cut having permanently lowered operating costs—or merely a cyclical bounce off the 2023 trough that leaves the company far below acquisition-era scale. ³ The record cannot presently distinguish these, because the decisive evidence—audited 2025 financials and full-year ad-spend data—is not public. The honest verdict is a stalemate on trajectory atop a settled finding of contraction.