The AI Bubble Is Popping; We Just Don't Know It Yet
Big tech's Q2 earnings revealed signs that the AI bubble may be popping, with investors reacting wildly to capex spending and free cash flow declines. The Register's podcast The Kettle breaks down what these numbers mean for investors and IT teams.
Key Earnings Signals
Editor-in-chief Matt Rosoff notes that massive-cap tech stocks are now swinging like penny stocks. Apple dropped 10% in a day after warning about component prices. IBM lost more value in one day than it had since Black Monday 1987. These are not normal market moves.
Meta's Capex Squeeze
Meta's free cash flow collapsed to under $1 billion this quarter, down from $8.5 billion the previous year. The cause: heavy capex on data centers for AI. Investors sold off on the news. Rosoff suggests Zuckerberg is betting that AI will improve ad targeting enough to justify the spend, and that excess capacity could be rented later—but that's speculation.
Amazon's Margin Games
Amazon's stock jumped 15% despite rising capex, driven by AWS growth and margins. However, analyst Corey Quinn pointed out that some of that margin came from hedging on energy prices—a one-time win, not core business. Investors need to separate such treasury tactics from real performance.
Advice for IT Teams
The hosts offer a practical warning: don't go all-in on frontier lab products across the enterprise during this volatile market. The uncertainty suggests a more cautious approach to AI adoption.
For the full discussion, including deeper analysis and more specific numbers, check out the episode.
📖 Read the full source: HN LLM Tools
👀 See Also
Georgia Court Order Contains AI-Hallucinated Legal Citations
A Georgia Supreme Court appeal revealed a trial court order contained at least five citations to nonexistent cases and five more to cases that don't support their cited propositions, with the prosecutor's proposed order containing the same errors.
OpenClaw agent spending patterns and lack of spending caps
A developer tracked OpenClaw agent spending over two months and found most agents average $40–$80/month in API and service charges when left unchecked, with spikes occurring on weekends and overnight. The default behavior is unlimited with no built-in spending cap.
Sora AI Video Economics: $20 User Costs OpenAI $65 in Compute
OpenAI's Sora AI video generation app reportedly costs $65 in compute per $20/month user, with peak inference costs estimated at $15 million daily versus $2.1 million total lifetime revenue.
The Vibe-Coding Noise Floor: How AI Slop Is Strangling Developer Communities
rmoff rants about the steady influx of low-effort AI-generated content in dev communities, from pointless GitHub repos to ghostwritten blog posts, and why it's driving away organic participation.