Yesterday, we talked about the “diffusion” of AI.  As Nvidia founder Jensen Huang suggested in his recent CES keynote presentation, not only will AI touch every part of the economy, it will discover new frontiers.

With that, we’ve been waiting for Wall Street to begin pricing in the reach of AI, and the benefits to the entire stock market, beyond just big tech. 

This chart below clearly shows the disparity.     

This is the Nasdaq and the Russell 2000, indexed at 100, from the date of the ChatGPT launch. 

Nasdaq futures have done 29% annualized since — almost three times better than small caps, almost three times better than the equal-weighted S&P 500.  

Moreover, the S&P 500 is on record highs, but a quarter of the constituent stocks have a negative three-year return (the post-“ChatGPT moment” era). 

Clearly the “Mag 7” stocks (NVDA, GOOG, META, AAPL, TSLA, AMZN, MSFT) have been priced as if AI would make everything else irrelevant.

That goes against Jensen’s AI “everything, everywhere” vision.

To start 2026, the wealth is indeed beginning to spread (so far).  The equal-weighted S&P 500 is outperforming the equal-weighted Mag 7 by 4 percentage points to open the year.

Is it because of policy tailwinds?  Yes (favors value).

But it’s also because AI adoption is driving a productivity boom, and that drives higher potential economic growth. 

That said, it’s reasonable to think that some companies will be disrupted — AI will be existential.  But it’s also reasonable to think that the rising tide of the economy will float all boats, AND that AI will outright transform some companies — turning the old into new.

We’re seeing it in data storage.  What was considered “legacy tech” just a year ago, is now in the cross-hairs of the AI infrastructure boom.

The two top performers in the S&P 500 today were Western Digital and Sandisk (up 17% and 28% today, respectively).  We own both in our Billionaire’s Portfolio.

This followed Jensen’s discussion yesterday on Nvidia’s new data storage architecture, to answer the intensive demands for inferencing (thinking, holding long context and massive libraries of data in real time).  This highlighted the insatiable and exponentially growing demand for flash storage — which Sandisk will supply.

But we already heard this in the most recent earnings calls from both Western Digital and Sandisk.  The message: the AI models (and the meters) are running non-stop.  Demand is high, and the data storage companies are effectively sold out.

And the amount of data storage required will scale in parallel with the amount of data that AI produces (which Jensen says is running 5x per year).