SpaceX, Adding It Up: The $235 Billion Cash Gap

Transcript

Scene 1: the number, fast. Two hundred thirty-five billion dollars. That is the gap between the cash SpaceX has said it will need through 2030 and the cash its public offering brings in. Fully issued, that offering meets about a third of it. It is the number behind the price on the share, a hundred and thirty-five dollars, and it is enormous: nearly a million dollars for every mile to the Moon. Here is the surprising part. The offering document holds every piece of this number and leaves the adding-up to the reader. Its own statement of the need runs about thirty words and stops short of a figure. So we did the adding, across five hundred pages and a good many more. Here is how.

This is that document, the registration statement, and the final version three weeks on, the same pages now carrying a mark, a planet, and the price. Nearly unreadable, so unread. I'm Greg Collins, at Cape Fear Advisors. Others have reached for this number too, and most valuation models simply assume it gets met. At this size, on a four-year clock, that assumption is worth stating and testing. My aim is the part underneath: the walk, how the number is built and checked, one filing against another, each figure traced to the filing it came from. This film is a test of that walk, one number at a time. So, the walk.

What a gap is

Scene 2. A gap has two sides. On one side, commitments: the cash and equity the company has already agreed to deploy. On the other, the money coming in, what the offering raises and what the business generates on its own. Line the two up in the same currency, and the gap is whatever the first side exceeds the second.

How a cash need is usually met

Framing. One word on what is normal. Usually, an offering meets or exceeds the cash a company says it will need. More recently, the hyperscaling businesses lean on future sources, capital to be raised later, once a milestone, a revenue figure or a scale, shows the model is on its way to profit. SpaceX sits on the line between the two: a capital-intensive business, where the raise funds the build, and a hyperscaling one, where a known risk is that more capital will be required. Most models make that assumption and move past it. We are going to state it, and test it.

Uses, line by line

Part 1. Start with what the money is for. Every item is disclosed, and each sits in a different filing. The bridge loan, twenty billion, in the debt footnote. The Anthropic compute build, ninety to a hundred billion, folded into segment capex. Cursor, sixty billion, in subsequent events. Terafab, fifty-five billion, filed with the county two weeks before the S-1 called it uncommitted. Valor, twenty billion, in the related-party note. Spectrum, eleven and a half billion. xAI's operating losses, thirty to forty billion, from history with no projection. Starship, called significant and left unpriced. In one currency, about two hundred thirty-five billion.

Sources, and their catches

Part 3. First, equity. The offering was the largest in history, about eighty-six billion fully issued. Set against the uses, it covers about a third. Then debt. In June they priced twenty-five billion in notes, a real source. But it arrived with a new use: to earn the investment-grade rating, they agreed to hold twenty-five billion in cash. The raise and the reservation cancel. Then the third source companies usually lean on: their own profit. Even the bullish models show losses through 2030, so there are no earnings to draw on, and the filing commits what it earns back into the build. Profit stays a future hope, and the bar does not move. Then Cursor, already sixty billion of the uses. They can settle it in stock, which does not shrink the gap; it pays it in dilution. And notice the one number doing both jobs: the share price is set once, in the offering, and then used, to raise and to settle. The same hundred and thirty-five dollars is the source and the currency.

What closes the rest

Part 4. So what closes the rest? The filing lists the main options. Issue more equity, which dilutes. Issue more debt, which they plan to do. Or spend less. There are also the edge moves companies reach for: a future customer can commit revenue, or even prepay; a vendor can extend flexible terms. Those are real, and at this scale they move the margin, not the middle. Which leaves spending less as the one clean answer, and it is the one they have ruled out, in their own words: capital expenditures will scale as quickly as they are able, funded by a range of debt and equity. The estimate is not a guess about the future. It is the size of a promise already made.

What a share buys

Part 5. One last thing, about the mission on the cover. The filing opens with it: to make life multiplanetary, and to extend the light of consciousness to the stars. That is the reason the company exists. Here is what the offering says the money is for: expanding AI compute infrastructure, launch capacity, and the satellite network. Last quarter, of about ten billion dollars in capital spending, more than seven went to AI compute, and about one to launch. Launch includes Starship, which is the mission's vehicle, so some of the capital does point that way. The mission is the north star, and in this offering it is the aspiration on the cover, not a funded line. The one place it sits near the money is the risk factors, as work that could be delayed or cancelled. We measured this number in miles to the Moon, because that is how a figure this large is felt. What the money buys is the build, and the build is mostly compute. The stars are what a share hopes for, later and in addition, not what it buys today.

Figure ledger

Every figure used in the video, its value, its label, and its source.