Every StaxInvesting article tagged artificial intelligence · 5 posts.
5 articles
This morning the question was whether the market punishes capital spending itself or only capital spending whose return cannot be measured. Intel answered it after the bell: fastest revenue growth since 2011, capex raised above $20 billion, and the stock up anyway. The distinguishing variable is not how much you spend — it is whether customers have already committed to buy what the spending produces.
Yesterday the question was whether AI demand was deferred or destroyed. The answer is neither. Alphabet delivered the clearest monetization evidence anyone has produced — 82% cloud growth, operating margin from 20.7% to 35.6%, a $514 billion backlog — and the stock fell anyway, because free cash flow went negative while spending accelerated. That is a different and more durable problem.
Alphabet, Tesla, and IBM report after today's close with no economic data to compete for attention. The tape is pricing one question: whether AI monetization justifies capital spending on track to roughly triple by 2027. This morning delivered a warning shot — a second enterprise software company flagging AI-related purchase delays — and the distinction that matters is whether that demand is deferred or destroyed.
On a Friday when semiconductors fell into a bear market on worries about AI capital spending, Meta was reported to be in early talks to lease out computing power. That looks like a contradiction and is actually the market's own answer to its own complaint. Among the big four hyperscalers, Meta is the only one with no public cloud — no channel to resell a single GPU-hour — which is exactly why its capex draws the most scrutiny. Here's what a compute-leasing business would change, and how to price it.
Overnight, Moonshot's Kimi K3 — a Chinese open-weight model rivaling US frontier systems — sent AI and semiconductor stocks into a third straight day of selling and revived the DeepSeek moment of 2025. The striking part isn't that chips fell; it's that a single competitive-model headline can reprice a trillion-dollar basket at once. This is why the AI trade behaves as one correlated bet, what happens to correlations when the thesis is threatened, and how to think about risk when diversification inside a theme quietly disappears.
