As I write this today, I can see the divergence in a way I can barely ever remember. In particular, trends related to what AI is doing to the Nasdaq. Look today’s price action:

The Dow is currently up about 0.1%

BUT…

The Nasdaq-100 is down almost 2.5%.

That is not a normal market message that you see often, not weekly, monthly, and very rarely yearly. That is a structure message because of underlying technological shift happening – money is flowing to new areas, TAM is changing, and ambitious pursuits like exploring outer space and building new physical tech, not software, such as robots, are taking center stage.

The market starting to ask whether the biggest tech companies in the index can keep spending as if they were infrastructure companies while still being valued like peak-margin software monopolies.

The mega-cap AI trade has a strange problem: If they spend, margins get pressured. BUT – if they do not spend, OpenAI, Anthropic, and the next generation of AI-native companies will take over.

So they spend: Data centers. Chips. Power. Talent. Models. Infrastructure. Compute. More compute. More capex.

That may be strategically correct.

It may also be financially painful for the next 5 years and they have no other choice. I am not even sure indexes like the Nasdaq-100 are fully understanding this.

A few historical reminders:

• In 2011, Nasdaq had to run a special rebalance because Apple had grown to roughly 20% of the index.
• After that rebalance, Apple’s weight was cut to roughly 12%.
• In 2023, Nasdaq had to do another special rebalance because the largest tech stocks had become too dominant again.
• Before the 2023 rebalance, the seven largest Nasdaq-100 stocks were about 56% of the index.
• After the rebalance, they were reduced to about 44%.
• Microsoft went from roughly 12.8% to 9.8%.
• Apple went from roughly 12.1% to 11.5%.
• Nasdaq’s own methodology includes concentration limits: no single company is supposed to exceed 24%, and the combined weight of companies above 4.5% is closely controlled.

The names change, but the behavior does not:

• In 2000, it was internet infrastructure.
• In 2011, it was Apple.
• In 2023, it was the Magnificent Seven.

Today, it is AI.

What does it cost to win AI? Because if everyone owns the same companies, and those companies are entering a multi-year spending cycle, then a lot of portfolios are quietly making the same bet: Peak-margin tech will stay peak-margin tech while spending like industrial infrastructure.

Maybe that works.

Maybe it does not.

So, the Nasdaq-100 is not just moving. It is rebalancing expectations and if you’re a trader or investor, the potential is everywhere as the index has no other choice but to continue rebalancing, de-risking, and reducing exposure to the same set of “typical” prior cycle tech/software winners.