The company generated more than $18.5 billion in revenue in 2024 but splurged nearly $21 billion on capital expenditure. Under GAAP accounting, it posted a net loss of close to $5 billion. Investor roadshows are expected to begin the week of April 20, with an S1 filing likely in late May or early June and an IPO targeting mid-June.
The numbers behind the headline
Strip out the GAAP losses and the picture is more complicated. Starlink, the satellite internet business, generated more than $7 billion in adjusted EBITDA last year. The launch business added nearly $1 billion on the same basis. Those two divisions are carrying the company.
The drag is xAI, the artificial intelligence unit Elon Musk acquired and folded into SpaceX. It lost between $1 billion and $1.5 billion. The majority of its revenue appears to come from X, the social media platform Musk bought in 2022, rather than from selling AI models commercially. That makes the xAI revenue story harder to tell to institutional investors who will want to understand the business on its own terms.
Cutting costs before the roadshow
SpaceX has already moved to address the burn. Michael Nichols, the executive who runs Starlink, has been appointed to oversee xAI as well. Other SpaceX executives are being brought in with an explicit brief to reduce costs and improve margins before the company goes public.
The logic is straightforward. Starlink is profitable and understood. xAI is neither. Giving Nichols oversight of both signals that SpaceX intends to run xAI the way it runs Starlink: capital discipline and a focus on the path to profitability rather than growth at any cost.
What investors are being asked to believe
The valuation being discussed sits between $1.5 trillion and $2 trillion. At that level, investors are not buying a profitable company. They are buying Musk's capacity to build one.
The investor base is expected to split roughly along those lines. Some will pay a premium to hold shares in a company Musk controls, in part because of the speculative possibility that SpaceX and Tesla could eventually converge. Others will look at the GAAP losses, the xAI drag, and the capital expenditure running ahead of revenue and conclude the valuation is too rich.
Shareholders also need to sell significant volumes of stock into the market to generate liquidity. That adds pressure to the pricing and to the roadshow.
The Musk question
Reporting from The Information has framed the Musk factor as a feature rather than a bug, noting that he has consistently moved forward despite obstacles that would have stopped others. That framing will appeal to some investors and alienate others.
Roadshow visits are scheduled at SpaceX's facilities in Texas and Tennessee, covering launch operations and data centres. The pitch will be made in person. Whether the valuation holds will depend on which kind of investor shows up.