--- title: "The Capex Bill Arrives Before The Revenue Does" type: "Topics" locale: "en" url: "https://longbridge.com/en/topics/100000000839811.md" description: "There is relatively little argument that artificial intelligence will eventually generate enormous economic value. However, the market spent Wednesday reminding everyone that "eventually" and "this qu..." datetime: "2026-08-06T06:02:30.000Z" locales: - [en](https://longbridge.com/en/topics/100000000839811.md) - [zh-CN](https://longbridge.com/zh-CN/topics/100000000839811.md) - [zh-HK](https://longbridge.com/zh-HK/topics/100000000839811.md) author: "[Ethan1](https://longbridge.com/en/profiles/24568761.md)" generator: "portal-rs" --- # The Capex Bill Arrives Before The Revenue Does There is relatively little argument that artificial intelligence will eventually generate enormous economic value. However, the market spent Wednesday reminding everyone that "eventually" and "this quarter" are two different balance sheets, and that they do not belong to the same shareholders. ### One conference call, two opposite reactions On its first public earnings call, SpaceX reported revenue of 7.81 billion dollars and an adjusted loss of 0.09 dollars per share. The number that mattered was further down. Quarterly capital expenditure came in above 18 billion dollars, roughly double the prior quarter, and 15.8 billion of that went to artificial intelligence infrastructure against 749 million dollars in the same quarter a year ago. On that same call, Elon Musk said SpaceX had decided to build exclusively on Nvidia's Vera Rubin NVL72 architecture. The tape did the arithmetic in real time. SpaceX fell about 13.6 percent. Nvidia rose for a fifth consecutive session. This is not irrational. It is the market correctly identifying who writes the cheque and who cashes it. ### The consensus framing is incomplete The popular reading of Wednesday is "AI bubble fears." I would suggest that is lazy. A bubble bursting looks like the whole complex selling together. What actually happened was a transfer of value within the same supply chain, from the buyer of compute to the seller of compute, executed inside a single hour of commentary. That distinction matters enormously for positioning. If you believe this is a bubble deflating, you reduce exposure to the entire theme. If you believe it is a margin transfer, you do something quite different, which is to ask where in the chain the cash is actually accruing and to weight accordingly. The historical parallel most people reach for is 1999. The more useful one is the telecom buildout of 1998 to 2001, where the equipment vendors booked spectacular revenue for several years while the carriers doing the building destroyed a remarkable amount of shareholder capital. The vendors were not wrong about demand. The carriers were not wrong about demand either. They were simply on the wrong side of a capital cycle, and the capital cycle does not care about being directionally correct. ### The specific risk sitting on today's calendar There is a second variable that has nothing to do with fundamentals. Approximately 911.5 million SpaceX shares, worth somewhere near 116 billion dollars, become eligible for insider sale from today, with insiders permitted to sell up to the first 20 percent of that eligible stock. Musk and a select group of insiders remain locked until the middle of 2027. What we know, with almost absolute certainty, is that supply of that magnitude does not clear without a price concession. What we do not know, and what nobody posting a target this week knows either, is how much of it actually comes to market. Lockup expiries are one of the few events where the range of outcomes is genuinely wide and the base rate is genuinely unhelpful. ### Portfolio implications Three points for readers positioning around this. First, separate the compute buyers from the compute sellers in your own holdings and look at the aggregate. Most retail portfolios I see are quietly long both sides and therefore have no view at all, which feels diversified and is not. Second, treat a rising capital intensity ratio as the warning it historically has been. A company that doubles capex in a single quarter while losses widen is making a claim about the future that will take years to verify. That is a legitimate thing to own. It is not a legitimate thing to own accidentally. Third, resist sizing anything around the lockup. Event-driven supply is the one situation where the disciplined position is a smaller one, not a cleverer one. ### What would invalidate this If SpaceX's next two quarters show that the 15.8 billion dollars is converting into contracted, recurring revenue at a rate the market did not expect, then the capital cycle framing is wrong and this is simply an underwriting of growth that happened to be front-loaded. I would want to see it in the revenue line, not in the commentary. The bottom line: the AI capital cycle is now visible enough to be measured, and measurement is usually where the enthusiasm phase ends and the differentiation phase begins. Own the chain deliberately, not by accident. ### Related Stocks - [NVDA.US](https://longbridge.com/en/quote/NVDA.US.md) - [SPCX.US](https://longbridge.com/en/quote/SPCX.US.md) - [TSLA.US](https://longbridge.com/en/quote/TSLA.US.md) - [NVDL.US](https://longbridge.com/en/quote/NVDL.US.md) - [07788.HK](https://longbridge.com/en/quote/07788.HK.md) - [07388.HK](https://longbridge.com/en/quote/07388.HK.md) - [NVDY.US](https://longbridge.com/en/quote/NVDY.US.md) - [NVDD.US](https://longbridge.com/en/quote/NVDD.US.md) - [NVDX.US](https://longbridge.com/en/quote/NVDX.US.md) - [NVDQ.US](https://longbridge.com/en/quote/NVDQ.US.md) --- > **Disclaimer: This article is for reference only and does not constitute any investment advice.**