The current AI market is like hot, moving fat in a skillet—fluid and competitive. The key strategic question is predicting when "the heat comes off and then everything's fixed." This "congealing" moment will lock in market leaders and make disruption much harder, marking the end of the wild early phase.

The AI era is not an unprecedented bubble but the next phase in a recurring pattern where each new computing cycle (mainframe, PC, internet) is roughly 10 times larger than the last. This historical context suggests the current massive investment is proportional and we are still in the early innings.

Current AI investment patterns mirror the "round-tripping" seen in the late '90s tech bubble. For example, NVIDIA invests billions in a startup like OpenAI, which then uses that capital to purchase NVIDIA chips. This creates an illusion of demand and inflated valuations, masking the lack of real, external customer revenue.

In an environment characterized by a series of sector-specific bull runs (e.g., from semis to metals), a winning strategy is to actively trade breakouts as they occur. This capitalizes on rotational leadership and momentum rather than relying on a static portfolio.

When a new technology stack like AI emerges, the infrastructure layer (chips, networking) inflects first and has the most identifiable winners. Sacerdote argues the application and model layers are riskier and less predictable, similar to the early, chaotic days of internet search engines before Google's dominance.

In a technology boom like the AI trade, capital first flows to core enablers (e.g., NVIDIA). The cycle then extends to first-derivative plays (e.g., data center power) and then to riskier nth-derivative ideas (e.g., quantum computing), which act as leveraged bets and are the first to crash.

The comparison to the dot-com bubble is incomplete. The current AI hype cycle hasn't yet been fueled by low interest rates or widespread leverage—factors that drove the final mania phase of the 1999 bubble. This suggests the market could get 'a lot crazier' before a significant correction.

Historical technology cycles suggest that the AI sector will almost certainly face a 'trough of disillusionment.' This occurs when massive capital expenditure fails to produce satisfactory short-term returns or adoption rates, leading to a market correction. The expert would be 'shocked' if this cycle avoided it.

The current AI market resembles the early, productive phase of the dot-com era, not its speculative peak. Key indicators like reasonable big tech valuations and low leverage suggest a foundational technology shift is underway, contrasting with the market frenzy of the late 90s.