Summary: Bubbles are a normal, human part of every big technology wave: railways, cars, credit cards, email, the dot-com internet, and now AI. Investor expectations rise for a reason that is partly justified, then overshoot, and anything with the hot label attracts money out of fear of missing out. The 2026 AI boom shows the signs: heavy, often debt-funded capex with little profit yet, extreme index concentration, and valuations last seen in the dot-com era. But a bubble can run for years before it pops, the underlying technology still raises productivity for almost every company, and for a long-term index investor a crash is a chance to buy the same stocks cheaper, not a reason to stay out.
Key ideas
- Since ChatGPT’s launch, AI stocks account for ~75% of S&P 500 returns, ~80% of earnings growth, and ~90% of the growth in capital investment (lazy-investing—111-the-ai-bubble, tamir).
- Big tech invests huge, partly leveraged sums in AI with little profit so far, because not investing means being left behind the way Nokia was by Apple; that leverage on high development costs is genuinely risky and could end in a burst (lazy-investing—111-the-ai-bubble, tamir).
- Fast-moving tech makes past investment obsolete: what a team learned two or three years ago can become irrelevant when a new model ships (lazy-investing—111-the-ai-bubble).
- Bubbles recur with every productivity leap (steam engine and railways, Fordism, credit cards, email), and the internet was a bubble in 1999, yet investing in 1999 was still an excellent long-term decision because the internet ended up lifting all 500 companies in the index (lazy-investing—111-the-ai-bubble, tamir).
- You don’t need AI ETFs to benefit from AI: “boring” companies (retailers, grocers, dividend aristocrats) adopt it too; the host’s six-person company now writes a landing page in five minutes instead of paying ~₪7,000 (lazy-investing—111-the-ai-bubble, tamir).
- A long-horizon investor (20+ years) should want a bubble to burst, to buy the same stocks cheaply (lazy-investing—111-the-ai-bubble, tamir).
Disagreements & open questions
- Is the AI boom a bubble? The host says it’s not yet proven: it may be the early days of a decade-long rally, or it may pop next year (lazy-investing—111-the-ai-bubble).
Takeaways
- Get AI exposure through broad indexes rather than AI-themed or leveraged AI ETFs (lazy-investing—111-the-ai-bubble)
Related
cape-ratio · market-timing · concentration-risk · fomo · index-investing