Quick Answer
As of 2026, X generates approximately $2.9bn in annual revenue — 35% below Twitter’s pre-Musk peak — with advertising still accounting for 68% of income despite a catastrophic advertiser exodus. The platform has since been folded into SpaceX via xAI in a deal valuing the combined entity at $1.25 trillion, fundamentally redefining what X actually is. Musk’s $44bn social media bet has quietly become something far more strategic — and far harder to value — than anyone who covered the original acquisition understood.
EBM Exclusive Take: The $44bn Infrastructure Play
The framing of X as a failed social media acquisition has always missed the point. Musk did not buy Twitter to fix Twitter. The evidence points consistently toward a single strategic objective: acquiring the world’s largest real-time human conversation dataset at the precise moment that data became the most valuable raw material in the global economy. That ad revenue fell 60% in the UK alone by January 2026 is not a crisis Musk is scrambling to fix — it is a distraction he has stopped pretending to care about. The advertiser exodus accelerated the pivot; it didn’t derail it.
The numbers that actually matter sit elsewhere. xAI is burning $28 million every single day to build the AI infrastructure that X feeds. X’s $2.9bn in annual revenue doesn’t exist to generate profit — it exists to fuel an engine whose ambitions are orders of magnitude larger. X is not the business. X is the data mine underneath the business. And that mine feeds directly into the $1.25 trillion SpaceX-xAI empire — an entity whose forthcoming IPO valuation depends in no small part on the quality and exclusivity of its AI training data. Every post, every conversation, every real-time signal flowing through X’s firehose makes that IPO case stronger.
Join The European Business Briefing
New subscribers this quarter are entered into a draw to win a Rolex Submariner. Join 40,000+ founders, investors and executives who read EBM every day.
SubscribeCritics are laughing at the empty storefront. Musk is quietly harvesting the gold mine beneath the floorboards.
Elon Musk bought Twitter for $44 billion in October 2022 and promised to turn it into an everything app. Nearly three years later, X has been absorbed first by his AI company xAI, then folded into SpaceX in a transaction valuing the combined entity at $1.25 trillion. The social network is now a subsidiary of a subsidiary — buried inside a corporate structure built around rockets and large language models. Understanding what X actually is in 2026 requires abandoning the framework used to evaluate it in 2022.
The Revenue Collapse in Full
The trajectory is worth stating plainly because the numbers are worse than most coverage has captured. In 2021, Twitter’s last full year as an independent public company, it generated $5.1 billion in revenue. By 2024 that had fallen to roughly $2.5 billion — a decline of more than 50% in three years. The culprit was an advertiser exodus triggered by Musk’s dismantling of content moderation, his public feuds with major brands, and his now-infamous instruction to departing advertisers at a DealBook conference to go and find themselves.
There was a partial recovery in 2025. X posted approximately $2.9 billion in total revenue for the year — its first year of growth since the takeover — driven partly by the return of advertisers who saw Musk’s post-election political capital as reason enough to re-engage. But even that improved figure sits 35% below what Twitter generated in its final pre-Musk year. As the broader collapse in digital advertising confidence has shown, platforms that lose institutional advertiser trust rarely recover it fully.
Musk himself acknowledged the reality in January 2025, stating publicly that user growth was stagnant, revenue was unimpressive, and the platform was barely breaking even.
Where the Money Actually Comes From
X’s revenue breaks into three streams, and the proportions expose the fundamental tension at the heart of the business.
Advertising remains the backbone, accounting for around 68% of total revenue — approximately $2.26 billion in 2025. Some major brands returned after Trump’s election victory, partly as a gesture of political alignment with Musk. But that momentum proved fragile: second-quarter 2025 ad revenue dipped 2.2% from the first quarter, suggesting the recovery was sentiment-driven rather than structural. Compared to Meta, which generated over $30 billion in ad revenue in 2025 alone, X is a rounding error in the digital advertising economy.
Subscriptions are the ambition that hasn’t landed. X Premium — offered across Basic, Premium and Premium+ tiers at $3, $11 and $22 per month respectively — generates an estimated $200 million annually. With a claimed 600 million monthly active users, that represents a conversion rate below 0.5%. Musk’s original investor projections called for 69 million paying subscribers by 2025. The actual figure is closer to 1.3 to 2 million — missing the target by a factor of roughly 35. Europe’s fintech leaders have demonstrated repeatedly that freemium conversion at scale requires a paid product that delivers tangible, recurring value. X has not built that product.
Data licensing is the quiet third stream and, in retrospect, the most strategically significant. Enterprises, institutions and developers pay for access to X’s real-time conversation firehose. This segment historically generated $500-600 million annually. Musk sharply restricted API access and raised prices, driving away smaller developers while making each remaining enterprise contract more valuable. The logic becomes clear when viewed through the xAI lens: X’s data isn’t just a revenue stream. It is the training substrate for Grok and future large language models.
The xAI Merger and What It Reveals
In March 2025, xAI acquired X in an all-stock transaction valuing X at $33 billion in equity — $45 billion including the $12 billion in acquisition debt — and xAI at $80 billion. The combined entity was valued at $113 billion on paper. Then in February 2026, SpaceX acquired xAI, creating a structure valued at $1.25 trillion that encompasses rockets, satellite internet, artificial intelligence and social media under a single ownership umbrella.
This was not a conventional acquisition. Both companies shared the same owner, overlapping investors and deeply intertwined operations. xAI’s chatbot Grok was already embedded across X’s interface. The platform’s billions of posts were already being used to train xAI’s models. The merger formalised what was already operational: X’s primary strategic value is no longer as a social network. It is as a real-time AI data pipeline — the largest and most current dataset of human public conversation on earth. As the race to control AI training data has intensified across the technology sector, that asset has appreciated dramatically in strategic value even as X’s advertising revenues deteriorated.
The Debt That Still Haunts It
None of this dissolves the financial reality of the acquisition structure. Musk loaded approximately $12 billion in debt onto X at the point of purchase, saddling the platform with more than $1 billion in annual interest payments. By November 2023, Fidelity had marked down its X stake to a valuation of just $5.3 billion — an 88% decline from the purchase price in barely a year. The rebound to the $33-44 billion range by early 2025 had more to do with Musk’s political capital than with X’s operational performance.
X’s adjusted EBITDA roughly doubled to $1.25 billion in 2024 compared to Twitter’s 2021 figure — but only because Musk cut approximately 80% of the workforce, reducing headcount from around 7,500 to fewer than 1,500. That is not a growth story. It is a cost extraction story on a platform whose revenue base is still contracting relative to its acquisition price.
What the Brutal Math Actually Says
The platform Musk bought for $44 billion generates under $3 billion in revenue, carries $12 billion in debt, pays over $1 billion annually in interest, and has a subscription product that missed its own projections by an order of magnitude. On those metrics alone, the acquisition looks like one of the most expensive miscalculations in corporate history.
But those metrics measure the wrong entity. X in 2026 is not a social media company trying to fix its ad business. It is a data infrastructure asset embedded inside the world’s most ambitious AI development programme, owned by the same person who controls the world’s leading private space company. Whether that convergence of AI, data and platform economics justifies the valuation is the only question that matters now. The jury remains out. But the question itself has changed — and that may be the most important thing Musk has achieved.
Related Analysis
