SpaceX formally filed its S-1 prospectus with the SEC on Wednesday, kicking off what could become the largest initial public offering in history. The company will list on Nasdaq under the ticker SPCX, and the filing offers the first detailed public look at both SpaceX’s business and the financial reality of Elon Musk’s AI venture, xAI.

The numbers are massive. SpaceX generated $18.67 billion in revenue in 2025, with more than $11 billion coming from its Starlink satellite internet service. But the company lost over $4.9 billion last year. That’s the cost of building rockets and satellites at scale while simultaneously betting billions on AI infrastructure.

xAI’s $6.4 Billion Burn

Buried in the IPO filing is something more revealing: xAI lost $6.4 billion in 2025. That’s not a typo. Musk’s AI company, which competes directly with OpenAI, Anthropic, and Google, burned through more than six billion dollars in a single year while building out its Grok AI platform.

The spending isn’t slowing down. According to the filing, xAI plans to buy $2.8 billion worth of natural gas turbines over the next three years to power its data centers. The company is already facing lawsuits over its existing generators, but that hasn’t stopped the expansion plans. SpaceX itself has set aside more than $500 million for potential litigation losses, partly to cover complaints that Grok created sexualized images. The IPO filing even lists Grok’s “spicy mode” as a specific risk factor.

This is what it actually costs to compete in frontier AI right now. While Anthropic just told investors it expects to hit profitability this quarter with $10.9 billion in revenue, xAI is still in the money-burning phase of building out compute infrastructure and training massive models.

The Musk Company Web

The filing also makes clear how intertwined Musk’s companies have become. A search for “Tesla” yields 87 mentions, xAI appears 356 times, and X shows up 267 times. Even smaller ventures like the Boring Company and Neuralink make appearances. Money and resources flow between these entities in ways that are difficult to track from the outside.

SpaceX’s willingness to absorb xAI’s infrastructure costs and litigation risks suggests the AI bet is central to Musk’s broader strategy. The $2.8 billion turbine purchase positions xAI to become a major cloud computing player, not just an AI model provider.

Nvidia Finds a New $200B Market

While Musk burns billions building AI infrastructure, Nvidia is printing money selling the chips that make it possible. The company posted another record quarter on Wednesday, though it warned that revenue growth would slow in the coming months.

CEO Jensen Huang isn’t worried. He says he’s found a “brand new” $200 billion market: CPUs for AI agents. As AI systems move from responding to prompts to actually taking actions autonomously, they need different hardware. Nvidia is positioning itself to dominate that transition the same way it dominated the training and inference chip markets.

The company also revealed it holds $43 billion worth of positions in AI startups. That’s venture capital at the scale of entire VC firms, embedded in a public chipmaker’s balance sheet. Nvidia isn’t just selling pickaxes in the gold rush anymore. It’s also betting directly on who strikes gold.

Meta Cuts Thousands to Fund AI

Not everyone is printing money or raising billions. Meta laid off thousands of employees this week in what management called a “continued effort to run the company more efficiently.” Translation: the AI investments are expensive, and the money has to come from somewhere.

The layoffs are meant to “offset the other investments we’re making,” according to an internal memo. Meta has spent tens of billions on AI infrastructure, model development, and AI-powered products across Facebook, Instagram, and WhatsApp. Those bets haven’t produced obvious returns yet, so the company is cutting costs elsewhere.

What It Means

The SpaceX filing pulls back the curtain on what’s usually hidden: the actual cost of competing in AI at the highest level. A $6.4 billion annual loss at xAI. A $2.8 billion bet on power infrastructure. A $43 billion portfolio of AI bets at Nvidia. Thousands of jobs cut at Meta to fund the arms race.

Anthropic’s path to profitability shows it’s possible to make the economics work, but it requires scale that most companies can’t reach. The gap between winners and everyone else is getting wider, and the cost of staying in the game keeps rising.

When SpaceX goes public, investors won’t just be buying into rockets and satellites. They’ll be buying into Musk’s bet that spending billions on AI infrastructure now will pay off later. The IPO will be a test of whether public markets are willing to fund that vision at the scale Musk wants.

If it works, SPCX could make Musk the world’s first trillionaire. If it doesn’t, the losses will be public, scrutinized every quarter, and impossible to hide behind private company opacity.

Either way, we’re about to find out what happens when AI ambitions meet public market accountability.