The 2027 AI Bubble: Even Bigger Than Dotcom?

Ticker Symbol: YOU•Sep 25, 2026

Summary

Alex provides a deep dive into the historical parallels between the 1999 Dotcom crash and the current AI investment surge. Alex notes that during the Dotcom bubble, companies built fiber optic networks for 12x growth but only realized 2x, leading to a 97% vacancy rate in cables and a 70% annual drop in bandwidth prices. In contrast, Alex highlights that today's AI traffic (tokens) is growing by 7x to 10x annually, while capital expenditure is only increasing by approximately 75%. This creates a supply-constrained market where demand significantly exceeds capacity.

Alex emphasizes that infrastructure prices for AI are rising rather than falling. For instance, the rental cost for Nvidia H100 chips has increased by 40% recently, and memory (DRAM) prices have nearly doubled. Alex also points out that the power grid is struggling to keep up with data center demand, with electricity contract prices in some regions increasing elevenfold. Alex concludes that while market skeptics like Michael Burry are shorting these stocks, the underlying data regarding utilization and revenue contracts suggests the bull market is supported by fundamental growth.

**Nvidia (NVDA):** Alex notes that Nvidia is currently the world's most valuable company but trades at a much more reasonable valuation than leaders of the Dotcom era. Alex points out that Nvidia trades at 28x earnings and a forward P/E of 18, compared to Cisco's 200x P/E in 2000. Alex argues that Nvidia's hardware is consistently at 98% utilization, proving that the demand for its chips is immediate and immense.

**Micron (MU):** Alex highlights that Micron's stock has more than tripled this year due to the skyrocketing demand for memory in AI servers. Alex explains that DRAM contract prices doubled in the first quarter of 2024 and continue to rise as supply cannot meet demand. Alex mentions that Micron trades at less than 9x next year's earnings, which Alex considers significantly cheaper than historical bubble peaks.

**Microsoft (MSFT):** Alex discusses Microsoft as a primary driver of the AI build-out with a massive backlog of future business. Alex notes that Microsoft has $678 billion in future business already under contract, representing an 84% increase year-over-year. Alex uses this figure to demonstrate that AI spending is not just speculative but is backed by long-term corporate commitments.