The AI Economy Is Creating a New Generation of Mega-Cap Companies

When a single company’s earnings report is about to be released and everyone somehow knows the entire week depends on it, a certain kind of silence descends upon a trading floor. On Apple days, that used to occur. Then it occurred during the Tesla era. These days, it occurs nearly every time Nvidia reports. The odd thing is that no one seems shocked anymore. As this develops, it’s difficult to ignore how much of the contemporary economy now depends on a small, closely knit group of names.

The numbers are approaching theatrical proportions. Last October, Nvidia surpassed $5 trillion, surpassing the GDP of all nations except China and the United States. Together, Microsoft, Alphabet, Amazon, and Meta intend to invest nearly $700 billion in AI infrastructure just this year. That isn’t a prediction. Even among analysts who have built their careers on bullish tech calls, that is a committed line item that appears in SEC filings and earnings calls. Although no one can agree on exactly what has changed, there is a feeling that something fundamental has changed.

The old valuation rules are being bent in strange ways. The company at the heart of this whole narrative, Nvidia, is trading at a forward price-to-earnings ratio that is not significantly higher than the S&P 500 as a whole. That in and of itself is odd. When compared to an index that is growing at twenty-four percent, you would expect a company that is growing its earnings at sixty-eight percent to appear pricey. It doesn’t. Even though doubters like Jamie Dimon and Ray Dalio subtly caution that some of this money will eventually be squandered, investors appear to think Nvidia has earned its position. It is possible for both to be true. Most likely, both are.

You can see the physical reality of all of this if you stroll around any data center construction site in Texas or Virginia these days. Beyond the horizon are concrete pads. Cooler boxes are used by crane operators to eat lunch. Behind chain-link, diesel generators hum. Just Google’s Midlothian facility is a forty-billion-dollar investment in Texas. In a slide deck, these are not speculative wagers. Concrete and copper are poured into them. In ten years, the buildings will still be standing, full of chips that someone, somewhere, will need, regardless of whether the business case warrants it.

The growing disparity between the mega-caps and the rest of the market is evocative of past periods in the history of the market. Cisco in 1999. The early 1970s Nifty Fifty. Years ago, Tesla faced similar skepticism and proved to be either a prophet or a cautionary tale, depending on who you ask. Nowadays, the dot-com comparison is frequently made, in part because the math encourages it. Approximately thirty percent of the S&P 500 is held by five companies. It’s the highest concentration in fifty years. In a winner-take-most economy, dominance might simply look like this. It might also have a negative outcome. To be honest, no one knows.

The second tier that is emerging behind the obvious giants is subtly fascinating. Due to memory chips, Samsung is getting closer to a trillion-dollar valuation. Because there is a shortage of HBM, which is necessary for every GPU on the planet to operate, Micron’s stock has increased by almost six hundred percent in just a single year. Few people fully understand the kind of monopoly that ASML, the Dutch company that manufactures the lithography machines that print the chips, enjoys. There are more than just one or two winners in the AI economy. Each tier is dependent on the one above and below it, forming an entire tiered hierarchy.

Observing the rising spending figures and declining free cash flow figures gives us the impression that we are in the early stages of an unidentified phenomenon. This year, Amazon is facing a negative free cash flow of almost seventeen billion dollars. Meta may spend up to $135 billion on capital expenditures. These businesses are not in financial ruin. Because they think falling behind would be more expensive, they are opting to spend this way. Perhaps they are correct. Perhaps, like railroads and oil, the AI economy is actually creating the next generation of long-lasting mega-caps. Or perhaps it’s something different, still developing, that we won’t fully comprehend until we look back.

The factories are operating for the time being. The chips are on their way. The funds are flowing. Everyone is aware of how the story might develop. Almost no one is prepared to abandon it.