Is AI a bubble? Maybe, maybe not. Who cares (greyenlightenment.com)

The piece argues that worries about an “AI bubble” are largely unhelpful: many experts predicted past bubbles that didn’t materialize, and the real pattern for transformative tech is long periods of heavy investment followed by a sudden “switch effect” where monetization (ads, subscriptions, services) turns prolonged losses into outsized profits. The author points to OpenAI’s rumored $500B valuation and parallels with Facebook, Amazon, Tesla and Uber—the latter showing concrete financial turnaround: Q2 2025 net income $1.4B, Adjusted EBITDA $2.1B, Gross Bookings $46.8B, EPS forecast rising from $0.63 to $3.49 by end-2026 and Q3 adjusted EBITDA of $2.19–2.29B—illustrating how seemingly frothy valuations can later be justified. For AI/ML practitioners, investors and founders the takeaway is pragmatic: VCs are pricing in long-term monetization and the possibility of a flip to profitability, so alarmist “bubble” calls offer little actionable guidance. There are exceptions—eg. Theranos—but these outliers don’t establish a universal outcome. The implication for employees (RSUs aside) and long-term investors is that market timing is risky; staying engaged with the sector may be more rewarding than trying to predict a pop. In short: froth alone doesn’t equal failure—structural product-market shifts and scalable monetization do.