#SpaceXRises9.8%AheadOfQ2Results 🦔SpaceX just posted Q2. Revenue $7.8 billion, up 92% year over year, and every segment beat estimates. Starlink is the only segment that made money, with 12 million subscribers and $1.66 billion in operating income. Space lost $542 million. AI lost $1.26 billion and spent $15.8 billion on capex in the quarter. The company also announced a $60 billion acquisition of Cursor. SpaceX ended the quarter with $100 billion in cash, but the market cap sits near $1.5 trillion, which puts the price at close to 50 times annualized revenue. My Take Two very different businesses live inside this earnings report. Starlink is one of the most impressive subscription businesses I've seen in years. Twelve million subscribers, up 100% year over year, with an operating margin near 40%. If SpaceX had gone public as a Starlink and launch business, this would be a straightforward buy at a reasonable valuation. The problem is that the other half of the company is a $16 billion a quarter capex furnace that just committed $60 billion more to buy a coding startup, and the market is paying 50 times revenue for the whole package. The Cursor bet only works if the developer surface holds value while the models underneath get cheaper. DeepSeek's release shows the models are already commoditizing faster than the deals SpaceX is signing to compete. Musk is spending $60 billion on the layer that sits closest to the customer, hoping that's where the moat ends up. I don't share that read. Coding is where models commoditize fastest, and enterprises route work to whatever's cheapest and works. At this quarter's capex pace, the AI business burns through the entire IPO cash pile in less than 18 months, so there's no room to be wrong. Starlink doesn't cover for that. Two different businesses, two different multiples, one stock price. $SPCX $TSLA $NVDAB