Read this assessment from an article yesterday. Given the widespread buzz around the issue, and having been directly affected by the **dot-com** bubble myself, I decided to run this through two leading AI models: **Claude Sonnet 5.5** and **ChatGPT Sol 5.6**. I also created a web page containing the prompt, a comparison table, vs table and the final verdicts.
The prompt for clarification:
"I read in an article that "AI Pricing is Not a Bubble" explained by: Valuation multiples for AI infrastructure stocks have fallen from a median forward P/E of 32 in April to 22 today, even as these companies continue to deliver strong results. Goldman Sachs argues this reflects strong pricing discipline rather than a speculative bubble. Now i want you to go out there and make deep research that confirm/not confirm this statement. do it by comparing it to the dot com (www) Bubble that heat traders. Make table of your finding. call it AI vs WWW Bubble and at the end summarize your verdict on AI Bubble. Test the statement as a hypothesis. is it clear?"
link to the full page by request.
Core Takeaways
* **Share Prices Are Not in a Dot-Com Bubble:** Both models agree that public AI share prices do not resemble the late 1990s dot-com crash. Public AI leaders are highly profitable, and the median forward price-to-earnings (P/E) ratio for AI infrastructure stocks dropped from **32x in April to 22x**, demonstrating market discipline rather than unbacked price speculation.
* **Bubble Risks Have Shifted to Capital Spending:** Both models warn that the true risk lies in massive capital expenditure (capex) on data centers, chips, power, and cooling. Hyperscaler AI investment is approaching **$800 billion to over $1 trillion**, with a growing portion funded by debt.
* **The "Earnings Bubble" Vulnerability:** Supplier profits currently make forward P/E ratios look reasonable. However, if AI revenue fails to reach the estimated **$300 billion annually** needed to break even on 2026–2027 spending, supplier earnings could plummet, exposing today's low P/E ratios as artificially deflated.
After page done and ready I asked **Gemini 3.6 pro** to read it and give its verdict:
public AI stock prices are not in a speculative bubble, but the massive infrastructure spending cycle carries a high risk of an earnings-driven correction.
To believe or not to believe?
NFA