Gavin Baker recommends AI investments in power, semiconductors, wafer capacity

Gavin Baker: The AI bubble debate is asking the wrong question

The issue isn’t valuations. It’s whether semiconductor investment has outrun real demand.

Historically, new tech cycles follow a pattern. Hot expectations, overinvestment, demand gap, collapse.

Today is different. Power and wafers are scarce. Capital can’t quickly build power plants, advanced foundries, or chips. Physical bottlenecks cap reckless supply growth.

Classic software scaled free after launch. AI recomputes answers on every query. The center of gravity shifts from code to power and manufacturing capacity.

What matters in AI investing isn’t the model. It’s the scarce inputs that stay scarce: electricity, semiconductors, wafers, and manufacturing chokepoints.

Direction isn’t enough. Volatility can knock you out. Fearing a bubble and missing structural demand is a mistake. Believing every AI company will win is a bigger one.