February 2026 saw $174 billion flow into AI and data investments in a single month. OpenAI alone raised $110 billion. Hyperscalers are projected to spend $610 billion on infrastructure this year. These are numbers that make financial media lose its mind, and they probably made you wonder if you're missing something.
You're not. Let me explain why, and then tell you the one thing that actually might matter.
The Boom Isn't Lifting All Boats
Here's the part the headlines skip: the S&P 500 is basically flat year-to-date as of late March. But underneath that calm surface, 74% of individual stocks have moved more than 5% in either direction. BlackRock calls that a 98th percentile dispersion event, comparable to the 2008 financial crisis in terms of how chaotic the underlying reshuffling is. Software stocks dropped 30% in 4 months. Semiconductors rose 30% in the same window. That's a 60-point divergence inside a single index.
Think of it like a grocery store where the total bill looks normal, but the price of eggs tripled and the price of bread collapsed. The average hides everything important.
What's actually happening is a rotation. Investors are selling companies that are spending heavily on AI infrastructure without clear profits, and buying companies that are already using AI to cut costs and grow margins. Meta cut 21,000 jobs while productivity went up. Microsoft says 35% of its code is now AI-written. Intuit and Salesforce are reporting 15-30% efficiency gains. The market is rewarding the adopters and punishing the builders who haven't shown a return yet.
So Does Any of This Change What You Should Do?
For most people: no. Not yet, anyway.
If you're contributing to a 401(k) or IRA through index funds, you already own the AI winners. You own Nvidia, Microsoft, Meta. You also own the software companies that got crushed. That's the deal with broad index investing: you get the whole messy picture, and historically that's still the better bet than trying to pick which slice wins.
The finance industry will absolutely try to sell you an AI-themed ETF right now. Some of those products are fine. Most of them charge higher fees to give you concentrated exposure to exactly the kind of dispersion that's currently eating software stocks. You're essentially paying extra to take on more risk at the moment risk is highest.
Fair point to the bulls: if you're already maxing your employer match, have 3-6 months of expenses saved, and carry no high-interest debt, adding a small position in a broad tech or semiconductor fund isn't crazy. The $2.2 trillion infrastructure pipeline through 2028 is real money going somewhere. I just wouldn't bet your financial foundation on guessing where it lands.
The tension I'll admit to: dispersion this extreme does sometimes signal a genuine structural shift, not just noise. Rick Rieder at BlackRock said AI has crossed from investment thesis to operating reality. He might be right. But "operating reality" and "good time to buy in" are two different things. The dot-com boom was also an operating reality in 1999.
What I'd actually do: check that your index funds are genuinely broad, not secretly tech-heavy. Some S&P 500 funds have 30%+ concentration in the top 5 stocks. If yours does, that's not diversification, that's a concentrated AI bet you didn't consciously make.
The $174 billion February number is real. The chaos underneath the flat index is real. The thing that isn't real is the idea that you need to restructure your savings to catch this wave. The wave is already in your portfolio. You just can't feel it yet.
