X Didn't Lose Its Advertisers by Accident. It Traded Them Away on Purpose.

In November 2023, on a stage at a New York investor summit, the owner of X was asked about the advertisers who had paused their spending on his platform. He did not court them. He did not apologize. He told them to go f**k themselves — and addressed Disney's CEO by name, who had spoken at the same summit earlier that day, adding: 'Hey Bob, if you're in the audience — that's how I feel.'⁹⁷ It was the rarest thing in business: a CEO saying out loud, to their faces, that he did not want his largest customers' money. Everyone treated it as a tantrum. It was actually a disclosure. He meant it, and the financials have been proving him right ever since.

The story everyone tells is that Musk's chaos accidentally scared the advertisers off — that better behavior could win them back, and the valuation would follow. That reading is comforting and wrong. The exodus was not a side effect. It was the deal.

He bought a brand-safety machine and unplugged the brand safety

Twitter sold one thing to make money: attention rented to brands that demanded their ad never appear next to something ugly. That is the entire premise of CPM advertising at scale — the advertiser pays a premium specifically not to be associated with the worst of the internet. Content moderation was not a cost center bolted onto the product. It was the product, the unglamorous machinery that let a cereal company feel safe spending money there. When Musk amplified an antisemitic post in October 2023, Apple, IBM, and Disney — the platform's 11th, 5th, and 8th largest advertisers — pulled their spend within days.⁷ They were not being squeamish. They were responding exactly as the business model trained them to. The machine worked as designed; he had simply removed the part that made it sellable.

That last line matters because it isn't a journalist's interpretation. It's X's own filing, in the dry compelled language of a UK statutory account, naming the cause itself. The same filing shows the bleeding: UK revenue fell from about $283 million in 2022 to roughly $92 million in 2023 — a 66% collapse — and kept falling to about $40 million in 2024.³ The UK is not a perfect proxy for the world, but the global picture rhymes. Independent estimates put X's worldwide ad revenue near $4.5 billion in 2022 and roughly $2.2 billion in 2023, with 2024 around $2 billion.⁵ Call it a $2 billion-a-year hole in the engine that paid for everything.

Why the valuation is a ratchet, not a correction

Here is the thesis, plainly: X did not suffer an advertiser exodus and then a valuation collapse as two separate misfortunes. It deliberately swapped its core monetization engine — brand-safety-dependent CPM advertising — for an ideological positioning that is structurally incompatible with that engine. The valuation isn't correcting toward some truer number. It's ratcheting, because each turn of the wheel makes the next recovery harder. Fidelity, an actual investor, marked its X stake down 78.7% by July 2024, implying a total value near $9.4 billion against the roughly $44 billion deal price.⁴¹ That is a fifth of what was paid. And the mechanism that keeps the ratchet from reversing is brutal in its simplicity: in August 2024, X sued the advertisers and their trade group for boycotting it, then expanded the suit in early 2025 to add Nestlé, Colgate-Palmolive, Lego, Shell, and Tyson Foods.⁸ You cannot win back customers you are also suing for the act of being customers somewhere else.

Watch the sequence and the logic is self-sealing. Brands pause over content; the owner insults them on stage; revenue craters; then the company sues the brands and the watchdog that coordinated their standards.⁷⁸ Each step is rational only if the goal was never to keep the CPM advertisers in the first place. If you genuinely wanted them back, every one of those moves was malpractice. If you had decided the platform's purpose was something other than a safe place to run a Tide ad, every one of them was on-strategy. The financials don't read like a company that lost its way. They read like a company that chose a different one and is paying the bill in public.

The honest objection: maybe the debt held, so maybe the rest can too

The fair counter is that the doom story has been wrong before. The roughly $13 billion in bank debt — the part Musk didn't cover from his own ~$33.5 billion equity check — was supposed to be a permanent albatross dragging the seven lending banks underwater.²⁶ It wasn't. By April 2025 the consortium had sold off the entire $13 billion to outside investors, largely escaping the trap.⁶ So if the debt overhang resolved, why not the advertiser story too? Because the debt and the revenue are different kinds of problems. Debt is a one-time block to be sold once appetite appears — and political proximity to a new administration helped manufacture that appetite — investor confidence in the debt sale was widely attributed in part to Musk's close relationship with President Trump.¹⁰ Advertising revenue is a recurring relationship that has to be re-earned every quarter from the very firms now named as defendants. One was a parcel to offload. The other is a flywheel that has to be restarted while its fuel is being sued. The debt got out. The CPM machine is structurally fenced off from its own customers.

Be careful with the easy villains, though. Musk has blamed the advertiser drop on pressure from an advocacy group, a self-serving and uncorroborated framing; the advertisers themselves cited brand safety, in their own words and in X's own filings.³⁷ And the headline numbers are noisier than they look — Fidelity won't disclose how it values a private security, and the often-quoted 'down 50%' figures range, in the careful estimates, from the mid-40s to the low-50s.⁴⁵ None of that softens the spine of it. The direction is unambiguous, the company itself names the cause, and the lawsuits foreclose the obvious cure.

A media business that sells brand safety and then dismantles brand safety hasn't made a mistake it can apologize its way out of. It has changed what it is. Musk paid about $44 billion for a machine that converted attention into advertiser money, and then spent two years proving he wanted a different machine entirely — and was willing to let the first one's revenue fall by half to get it.¹⁵ The valuation collapse isn't the market punishing chaos. It's the market pricing a swap that was made on purpose. The genius and the failure are the same act, seen from two sides: he got exactly the platform he wanted, and it is worth a fifth of what he paid for the one he gave up.