SpaceX and Meta Can Fund AI, but the Hard Part Is Paying It Back

The most important signal from the latest SpaceX and Meta announcements is not weakness in either company’s underlying business. It is the scale of capital required to pursue artificial intelligence. SpaceX is now financing Nvidia GPUs with $40 billion of debt, while Meta has created a new non-voting share class that can be used to raise equity as its AI infrastructure programme expands. The common thread is simple: AI is moving from a technology investment cycle into a financing cycle.

That distinction matters for investors because a company can be operationally stronger while its financial risk is rising. SpaceX reported explosive third-quarter growth, and Meta continues to produce tens of billions of dollars of quarterly operating cash flow. But both are committing capital at a pace that changes the relationship between operating performance, balance-sheet capacity and shareholder returns.

SpaceX’s $40 Billion Raise Is Bigger Than the Headline

SpaceX reported $14.6 billion of third-quarter revenue, up 161% year over year, alongside $4.4 billion of net income and $8.8 billion of adjusted EBITDA. Those numbers make the company’s borrowing capacity understandable. The problem is the spending rate: third-quarter capital expenditure reached $19.6 billion, while nine-month purchases of property and equipment reached $48.1 billion.

SpaceX had $38.0 billion of aggregate principal indebtedness at September 30, including $25 billion of senior unsecured notes issued in June. Musk has now said the company has secured another $40 billion of debt financing to purchase Nvidia GPUs, primarily for delivery in the first half of 2027. If the new financing is added without offsetting repayments, the principal debt load would roughly double from the September figure.

The important point is not that $40 billion automatically makes SpaceX overleveraged. It does not. SpaceX also had $51.3 billion of cash and $38.6 billion of short-term marketable securities at quarter-end. The issue is what the new debt is buying and how quickly those assets have to generate returns.

Earlier reports indicated a proposed structure of roughly $10 billion in bank loans and $30 billion in investment-grade debt. Subsequent reporting suggested some tranches could price around 7.75% to 8.3%. Those are reported figures rather than final terms, but they provide a useful sensitivity: $40 billion financed at an 8% average cash rate would imply roughly $3.2 billion of annual interest before fees and other financing effects.

That amount is manageable against SpaceX’s current earnings power if the AI business continues scaling. It becomes more important if compute capacity grows faster than monetization. The company is effectively borrowing against a future revenue stream that still has to be built, while its established Starlink and launch businesses provide the current earnings base supporting the credit structure.

SpaceX generated $13.1 billion of operating cash flow in the first nine months of 2026 but spent $48.1 billion on property and equipment. That gap was financed through a combination of debt, equity proceeds and balance-sheet management. The SEC filing says investing cash outflows increased by $71.6 billion from the prior-year period, primarily because of higher marketable-security purchases and a $34.2 billion increase in capital expenditure tied to data centers, infrastructure and launch facilities.

This is why the $40 billion financing should not be treated as a one-off transaction. SpaceX itself says future capital needs may vary significantly with the pace and scale of its growth initiatives and that additional financing may be sought through debt or other arrangements. The company is telling investors, in its own filing, that capital requirements remain a moving target.

SpaceX reported debt at September 30, 2026, plus an illustrative $40 billion addition. Source: DaveFinances.com

Chart 2. SpaceX nine-month operating cash flow versus property and equipment purchases. Source: DaveFinances.com

Meta Is Building a Financing Valve Before It Needs to Turn It On

Meta’s move is different because it does not need to borrow $40 billion simply to keep the lights on. Third-quarter revenue rose 26% to $64.6 billion and operating cash flow was $32.8 billion. Yet capital expenditure, including principal payments on finance leases, reached $34.9 billion, leaving free cash flow at negative $2.1 billion.

Meta expects full-year 2026 capital expenditure of $135 billion to $145 billion, and management says it will continue investing aggressively beyond 2026. The company ended September with $82.5 billion in cash, cash equivalents and marketable securities. So the immediate question is not liquidity. It is how long internal cash generation can finance an infrastructure programme of this size without materially changing the capital structure.

Meta’s new Class C shares are not simply a cosmetic stock split. The initial one-for-one distribution to existing Class A and Class B holders does not raise cash, but Meta explicitly says the new non-voting class can later be used for capital raising, acquisitions, employee compensation and other corporate purposes. The structure becomes effective in January 2027.

That creates a financing option without adding voting power to Zuckerberg’s control. It also creates the potential for economic dilution if Meta eventually sells substantial amounts of Class C stock. JPMorgan analyst Doug Anmuth has argued that the company should issue the shares sooner rather than later to remove the overhang, while other reports have pointed to possible financing in the tens of billions. Meta itself has not announced a specific offering.

Meta’s Cash Flow Is Already Telling Investors Where This Goes

The key number is the relationship between operating cash flow and capital expenditure. In the third quarter, Meta generated $32.8 billion from operations and spent $34.9 billion on capital expenditure. That is a small gap relative to Meta’s scale, but it is significant because the company is simultaneously guiding to $135 billion-$145 billion of full-year capex and warning that AI investment will remain aggressive.

Meta can fund a large portion of that spending internally if advertising revenue continues to compound. But the more AI infrastructure expands, the more valuable external capital becomes as a way to preserve flexibility. Equity avoids interest expense but dilutes economic ownership; debt preserves ownership but creates fixed obligations. The Class C structure gives Meta another lever before management has to make that trade-off.

Chart 3. Meta Q3 2026 operating cash flow, capital expenditure and free cash flow. Source: DaveFinances.com

SpaceX finished October 23 at $147.49, up 1.89% that day, even as investors absorbed the scale of the company’s infrastructure ambitions. Meta had a much harsher immediate reaction to the Class C announcement, with shares falling sharply in early trading. The difference is telling: investors can reward extraordinary growth while still demanding a higher return for taking financing risk.

Meta’s reported capital-structure change also arrives as credit markets are becoming more sensitive to AI spending. Reuters reported that Meta’s five-year credit-default swap spread widened to a record level in the aftermath of the announcement. That does not mean the company is approaching a credit event. It means investors are beginning to put a price on the possibility that the AI build-out will require more external capital than the market previously assumed.

The Bigger Risk Is Not Debt — It Is Return on New Capital

The cleanest way to think about both companies is to stop asking whether they can raise the money. They clearly can. The harder question is whether the incremental capital earns an adequate return after financing costs, depreciation and the risk that computing equipment becomes obsolete faster than expected.

SpaceX has an unusually broad revenue base through Starlink, launch services and its emerging Super Intelligence business. Meta has one of the strongest advertising engines in technology. Those businesses reduce the risk compared with a pure-play AI infrastructure borrower. But they do not eliminate it. The larger the capital programme becomes, the more investors need evidence that each additional dollar of AI infrastructure is producing incremental revenue, margins or strategic value rather than simply increasing capacity.

For SpaceX, the next useful disclosures are the final terms of the $40 billion financing, the maturity profile, interest rates and the amount of additional debt needed for data centers, satellites and other infrastructure. The operating metric that matters most is the conversion of 1.7 gigawatts of nameplate compute into contracted revenue and cash generation.

For Meta, watch the actual structure and timing of any Class C or other financing. A $20 billion equity offering has a very different effect from a convertible issue or a large debt deal. The other key variable is 2027 capex: if spending remains near or above the current trajectory while free cash flow stays weak, external financing will become increasingly difficult for investors to ignore.

The broader market implication is larger than either stock. The AI boom has already created extraordinary demand for chips, power, data centers and networking equipment. The next layer is the financing required to build all of it. SpaceX is using debt. Meta is preparing equity capacity. If that pattern spreads across hyperscalers and AI infrastructure companies, the cost and structure of capital will become a much more important part of the AI investment thesis.

Original Post