Yesterday, SpaceX started trading on the Nasdaq under the ticker SPCX. The company sold 555 million shares at $135 each, raised $75 billion in a single day, and landed a valuation of $1.77 trillion - making it the seventh most valuable company in the United States, ahead of Tesla, its founder's other listed company.

That's the biggest initial public offering - meaning the first time a private company sells shares to the general public - in the history of capital markets. Not by a little. The $75 billion haul more than doubled the previous record, set by Saudi Aramco in 2019, which raised roughly $25.6 billion.

And veteran short-seller James Chanos thinks this is precisely where the alarm bells should start ringing.

"Wall Street has a printing press too, just like the Fed," Chanos told Bloomberg Television on June 12. "It just takes a while to get going."

The size of the IPO is not the celebration, in his reading. It's the warning signal.

The background

Short-selling is a bet that a stock's price will fall. It works by borrowing shares, selling them immediately, waiting for the price to drop, then buying them back cheaper and pocketing the difference. It's a risky business - the stock can keep rising indefinitely, making the bet infinitely expensive. James Chanos is probably the most famous short-seller alive, best known for spotting the fraud at Enron before the company collapsed in 2001.

When Chanos talks about IPO waves, he has a specific thesis: history shows that whenever the volume of new share issuances - measured relative to the size of the broader economy - spikes dramatically, markets tend to fall shortly after. He compared 2026 to 1999-2000, the height of the dot-com bubble, and to 2021, the year of the meme stock and SPAC craze. A SPAC, or Special Purpose Acquisition Company, is essentially a shell company that raises money through a stock listing with the explicit purpose of then going out to acquire a real business - a shortcut for private companies to go public without the scrutiny of a traditional IPO process.

In both those prior episodes, a flood of new equity supply met a flood of investor demand, generating enormous paper wealth before the eventual correction. Chanos's view is that 2026 is shaping up to be larger than either.

The SpaceX IPO is only the beginning. According to NPR, the SpaceX listing is expected to be the first of a trio of mega-IPOs from AI companies this year, with OpenAI and Anthropic also reportedly preparing their own offerings.

What is actually happening

SpaceX is not a simple company anymore. In February 2026, the company completed a merger with xAI - Elon Musk's artificial intelligence startup - in a share exchange that valued the combined entity at $1.25 trillion. That means anyone buying SpaceX shares today is also buying Grok (xAI's AI model), the X social media platform (which was folded into xAI earlier), and the Starlink satellite internet service - all bundled into one.

The IPO prospectus revealed that SpaceX reported revenue of $4.69 billion in the first quarter of 2026, up from $4.07 billion a year earlier. For all of last year, revenue totalled $18.67 billion. The company posted a net loss of $4.28 billion in the most recent quarter, after losing $4.94 billion in 2025.

In other words: revenues are growing fast, but the company is spending considerably faster than it earns.

At $1.77 trillion in market capitalisation - the total value the stock market assigns to a company - SpaceX's price-to-sales ratio (what investors are paying for every dollar of annual revenue) is roughly 92. Chanos put the number closer to 110 times revenues based on his calculations at time of taping. For comparison, Tesla - itself considered expensive by most analysts - trades at around 14 to 20 times revenues. Morningstar analysts assessed SpaceX's fair value at $780 billion, less than half its IPO price, citing "very high" uncertainty and complicated governance under Musk, who simultaneously runs Tesla and other companies.

The Starship rocket, which underpins the entire longer-term vision of the company, has completed 12 test missions but has still not achieved stable Earth orbit. Half of those missions ended in some form of failure.

The money trail

The most revealing story in the SpaceX prospectus is not the rocket programme or even the Starlink subscriber count. It's a pivot buried in the filing that Chanos says dramatically changes what kind of company investors are actually buying.

Weeks before the IPO, the filing disclosed that SpaceX had signed a deal to rent out the entire computing capacity of the Colossus 1 data centre - the facility in Memphis originally built to train xAI's Grok models - to Anthropic. Under that arrangement, Anthropic committed to paying $1.25 billion per month, running until May 2029.

Then, on June 5, a separate filing revealed that Google had signed a deal to lease additional compute capacity at $920 million per month, running through June 2029.

Combined, the two agreements represent approximately $26 billion in annualised compute revenue from two of the best-funded AI companies in the world.

The business logic is straightforward: Colossus 1 was built to train Grok. Grok's monthly app downloads fell from more than 20 million in January to about 8.3 million by April, a 60% drop in three months, according to AppMagic data. xAI moved its training operations to the newer Colossus 2 facility, leaving Colossus 1's 220,000 Nvidia GPUs sitting largely idle - a pile of expensive hardware generating no returns.

Renting it out to Anthropic and Google was the logical response. But Chanos's point is that the nature of the business changed in the process. What was being pitched to investors as an AI model company - a high-multiple, high-growth technology firm - quietly became a data centre landlord.

"That's the NIO cloud model," Chanos told Bloomberg. "That's your equipment lessor. And that is a much lower valued business in the marketplace."

A NIO cloud, or neo-cloud, is an industry term for companies that buy computing hardware - usually Nvidia chips - and lease it out to AI companies that need processing power but don't want to own the infrastructure themselves. CoreWeave is the best-known example. These businesses are not software companies. They're more like industrial finance, with returns driven by lease rates and depreciation schedules rather than product development. The market values them accordingly.

The paradox is hard to miss. Three weeks before the IPO, the xAI division posted a $2.47 billion operating loss in Q1 2026, while AI capital expenditure reached $7.7 billion in the first quarter alone. SpaceX spent more on AI infrastructure in three months than the entire Saudi Aramco IPO raised in 2019. The deals with Anthropic and Google now provide revenue to offset some of that spend - but by transforming the company's business model just before listing.

Chanos flagged something else that rarely gets discussed: the broader accounting effect of technology capital spending booms. When a company like SpaceX buys chips from Nvidia, it capitalises that expense - meaning it spreads the cost over five to ten years on its books rather than recording it as an immediate loss. Meanwhile, Nvidia recognises the full sale as revenue and profit right away. During a spending boom, this accounting mismatch inflates reported earnings across the tech sector. When the boom ends and order books shrink, earnings collapse sharply. He noted it happened between mid-2000 and mid-2001, when S&P earnings dropped 40% over two years after rising 30% in the preceding two.

What people are doing about it

Among institutional investors, the demand for the SpaceX IPO was extraordinary. The order book ran five times oversubscribed, with the $75 billion raised exceeding the combined total of every other 2026 IPO so far.

The IPO prospectus confirmed that Musk will retain more than 80% of stock voting power, meaning no public shareholder can influence the company's direction regardless of how many shares they hold.

Retail investors - ordinary members of the public buying through brokerage accounts - faced limited access. SpaceX cut its retail allocation to the low-20% range, meaning the vast majority of shares were distributed to institutional buyers. Some brokerages reportedly lowered their minimum account balance requirement to $2,000 from $500,000 to allow broader participation.

On its first day of trading, SpaceX shares gained 19%.

Chanos, for his part, noted that SpaceX cannot currently be shorted - the mechanics of short-selling require that shares be available to borrow, and a newly listed company typically has very limited borrowable float in early trading. Whether the stock makes it onto his "radioactive" watchlist, as he described it, remains an open question.

For those already holding Tesla, the picture is complicated. Tesla shares were down more than 11% year-to-date heading into the SpaceX listing. Tesla disclosed a $2 billion equity investment in SpaceX and a semiconductor manufacturing partnership at Gigafactory Texas, creating an indirect exposure for existing Tesla shareholders - though not one they voted for.

Broader market observers point to what comes next: OpenAI and Anthropic are both reportedly preparing public offerings later in 2026. If they price at the valuations currently being discussed - OpenAI around $830 billion and Anthropic around $350 billion - the total volume of new equity issuance will dwarf anything seen since the dot-com era.

The bottom line

SpaceX raised more money in its IPO than any company in history. It is simultaneously a rocket manufacturer, a satellite internet provider, an AI model company that has lost market share, a data centre that now rents its idle computing capacity to competitors, and a bet on Elon Musk's ability to eventually make all of it worth $1.77 trillion. James Chanos's warning is less about SpaceX specifically and more about what the sheer size of this IPO signals: that the supply of new equity is meeting demand in the same way it did in 1999 and 2021, both of which ended badly. The story, for now, is flying. History suggests the landing matters.

Timeline

  • 2001 - Enron collapses after accounting fraud. James Chanos had publicly flagged concerns before the collapse, establishing his reputation as a short-seller.
  • July 2025 - SpaceX conducts a private secondary share sale valuing the company at approximately $400 billion.
  • December 2025 - SpaceX conducts another private share sale at $800 billion valuation. IPO plans reported to be advancing for mid-2026.
  • January 2026 - SpaceX and xAI merger talks advance, with two Nevada corporate entities registered on January 21 to facilitate the deal.
  • February 3, 2026 - SpaceX and xAI announce merger, creating a combined entity valued at over $1.25 trillion.
  • May 2026 - SpaceX files S-1 IPO prospectus with the SEC. Filing discloses Anthropic deal worth $1.25 billion per month to lease Colossus 1 compute capacity.
  • June 5, 2026 - SpaceX files separate disclosure of Google compute deal worth $920 million per month through June 2029.
  • June 11, 2026 - SpaceX prices IPO at $135 per share, raising $75 billion - the largest IPO in history.
  • June 12, 2026 - SpaceX begins trading on the Nasdaq under SPCX. Shares close up 19% on the first day. James Chanos appears on Bloomberg Television warning that the scale of issuance mirrors prior market peaks.

Summary

Who: Elon Musk's SpaceX, merged with xAI; short-seller James Chanos; investors including Anthropic, Google, and public market participants.

What: SpaceX raised $75 billion in the largest IPO in history, pricing at $135 per share and a $1.77 trillion valuation. Chanos warned publicly that the size of the offering, combined with anticipated mega-IPOs from OpenAI and Anthropic, echoes conditions that preceded previous market downturns.

When: The IPO priced on June 11 and began trading June 12, 2026. Chanos's Bloomberg Television appearance aired June 12.

Where: Nasdaq, New York. xAI's Colossus data centres are in Memphis, Tennessee. SpaceX operates globally through Starlink.

Why: SpaceX required public capital to fund ongoing AI and space infrastructure spending. The company's pivot - from an AI model company to a data centre lessor - changed the economic logic of the deal just before listing. Chanos's concern is structural: he argues that record equity issuance historically precedes sharp market corrections, independent of any single company's merits.