The Rhyme: Anthropic Files the Largest IPO Ever & Its 2012 Echo

"History doesn't repeat… but it rhymes." — Mark Twain

Anthropic is expected to publicly file its S-1 prospectus as early as today, clearing the last procedural hurdle before a November listing that could raise as much as $100 billion at a valuation near $2 trillion. That would make it the largest initial public offering in history, topping the $75 billion SpaceX raised four months ago. The filing would hit the SEC's website on the same morning that the Nasdaq sat 5% below its October 6 record high, with Meta down 7%, SpaceX down 7%, and the AI trade under the heaviest selling pressure since spring.

ANTH Target Valuation ~$2T  |  H1 Rev $16.3B (+15x YoY)  |  H1 Net Loss $80.7B  |  Compute Commitments $600B+  |  S&P 500 7,658  |  10Y 5.02%  |  Brent $111

The last time the defining technology company of a generation dropped a blockbuster S-1 into a market that was already losing its nerve, it was February 2012. The company was Facebook.

At 11:00 a.m. Eastern on Friday, May 18, 2012, trading was scheduled to begin for the biggest technology IPO the world had ever seen. Mark Zuckerberg stood in Facebook's Menlo Park cafeteria, surrounded by eight hundred employees in jeans and hoodies, staring at a screen. The plan was simple: he would press a button in California, the Nasdaq would open the stock in Times Square, and 421 million shares of Facebook would begin trading at $38 apiece. The plan broke immediately. A software glitch inside Nasdaq's system froze the opening auction. The seconds ticked into minutes. Eleven o'clock came and went. Traders in New York picked up phones that had nothing to tell them. Zuckerberg waited. The room in Menlo Park went quiet. It took thirty minutes for Nasdaq to switch to a failover system and push the stock into the market at 11:30.

When shares finally opened, they popped to $42.05 on a wave of pent-up orders and then started sinking almost immediately. By early afternoon, Facebook was sliding toward its $38 offer price, and Morgan Stanley's traders were buying shares by the tens of millions to keep the stock from falling below it. They succeeded, barely. Facebook closed its first day at $38.23, up twenty-three cents. The Nasdaq finished down 1.2% to 2,779, and the S&P 500 dropped 0.7% to 1,295, capping what was already the worst week of 2012. Europe's sovereign debt crisis was boiling over in Greece and Spain, and the biggest tech IPO in history had just been held together with duct tape and a bank's balance sheet.

Facebook had filed its S-1 on February 1, three and a half months earlier, with an initial placeholder target of $5 billion — a figure the underwriters would push to $16 billion by pricing day. The numbers looked impressive: $3.7 billion in 2011 revenue, $1 billion in profit. But buried in the risk factors was a warning that would haunt the stock for the next fourteen months. Facebook was making almost no money from mobile users, and mobile was where all its growth was going. By the time the company updated its S-1 on May 9, the underwriters had already started cutting their revenue estimates quietly, passing revised numbers to select institutional investors over the phone while retail buyers were still reading the original prospectus. The SEC would later fine Nasdaq $10 million for the technical debacle. The lawsuits piled up. And the stock kept falling.

By September 4, 2012, Facebook traded at $17.73, down 53% from the IPO price. The company that was supposed to prove the social media era could deliver returns to public shareholders had instead become a symbol of Silicon Valley overreach. Analysts wrote it off. Magazine covers asked if Zuckerberg had blown it. Then, quietly, something changed. Facebook cracked mobile advertising. It rebuilt its entire ad platform around the phone screen, and by mid-2013 mobile ads accounted for more than 40% of advertising revenue, up from essentially zero at the time of the IPO. The stock crossed back above $38 in August 2013, fourteen months after it first traded there, and it never looked back. A decade later, the company was worth over a trillion dollars.

The lesson was not that the IPO was a great buy. It was that the biggest IPO of a new technology era almost always arrives at exactly the wrong moment for the market, and exactly the wrong price for the stock, even when the underlying company turns out to be real.

Both companies filed their prospectus as the undisputed leader of a new technology wave, at the biggest valuation Wall Street had ever seen for a new listing, into a market that was already cracking. The defining company of an era does not make the IPO safe. It makes it a referendum.

Here is what actually happened after Facebook went public at the peak of the hype. The stock lost 53% in 109 days. It fell from $38 on May 18 to $17.73 on September 4, 2012, in a grinding slide that accelerated every time the company released numbers that confirmed what the S-1 had already warned about: mobile revenue was a question mark, and the market wanted an answer before it would pay a hundred-billion-dollar price tag.

The S&P 500 dropped roughly 10% from its April high to its June low, dragged down by the European debt crisis. But the broader market recovered by September. Facebook did not. The stock was alone in its misery because the problem was specific to its valuation, not the economy. Investors had paid a price that assumed perfection, and they got a company in the middle of a difficult transition.

The smart money did not buy on IPO day. It did not buy the first dip, or the second. It waited until the thesis was either confirmed or destroyed. The people who made fortunes in Facebook bought it in late 2012 and early 2013, after the company had proven it could sell mobile ads at scale, at prices between $19 and $28. They bought certainty, not hope. The fourteen-month gap between the IPO and the recovery was not dead time. It was the market doing what markets do: forcing the company to earn its price.

Anthropic's version of the mobile question is compute. Can it convert $600 billion in infrastructure commitments into durable revenue that grows faster than the bills come due? The S-1 shows $16.3 billion in first-half revenue against $9 billion in compute spending in the same period. That ratio works at this scale. Whether it works at ten times this scale, with noncancelable cloud contracts stretching out a decade, is the question that will take quarters to answer, not a single opening trade on a November morning.

The biggest IPO of a technology generation tends to mark the moment the era stops being a bet and starts being a stock. Facebook's IPO was the beginning of a 53% drawdown and a fourteen-month wait. The company was real. The price was not. Watch the first earnings report after listing, not the first trade. That is where the mobile question got answered for Facebook, and where the compute question will get answered for Anthropic.

Alphabet and Microsoft both report earnings on Wednesday, October 28, the same afternoon the Fed announces its rate decision. If either company signals that AI cloud revenue is decelerating while capex keeps climbing, it would echo the exact warning Facebook's revised S-1 buried on May 9, 2012, nine days before the stock went public and the floor gave way.