I want to talk about something that's been bothering me. Right now, you can park money in a 10-year US Treasury and earn about 4.6% a year. No drama, no drawdowns, just a quiet, guaranteed return. The S&P 500's earnings yield? Almost identical. Same number, completely different experience. One sleeps quietly in a drawer. The other can lose a third of its value before lunch.
So why does anyone bother with stocks?
Because stocks can grow. A bond pays you a fixed coupon and gives your money back. A stock gives you a claim on a living business, one that can earn more next year than it did this year, and more the year after that. The extra return investors demand for taking that ride instead of clipping coupons is called the equity risk premium. Right now it sits at about 5.4 points above Treasuries, which puts expected equity returns somewhere near 10% a year. Historically, that's normal. Not exciting, not worrying. Just the long-run average doing its thing.
But that premium has never been evenly distributed. The names that drive index returns decade after decade tend to be the ones riding the biggest technological wave of their era. Right now, that wave is AI.
When you buy the index, you are not buying "the market." You are placing a bet that a handful of companies, most of them now building AI into every corner of the economy, will keep compounding faster than everyone else. That is the real index bet, and it always has been.
If the equity risk premium has always been driven by the next great wave of technology, then what investors are actually pricing today is whether AI works. Not as a stock ticker or a trading theme, but as the force that reshapes how industries operate. Drug development that takes two years instead of ten. Supply chains that route themselves. Risk models rebuilt from scratch overnight. The companies doing this aren't all going to look like traditional tech. They'll be healthcare firms, logistics operators, banks. They'll just be the ones that figured it out before their competitors did.
Every generation gets one of these. Railroads, electrification, semiconductors, the internet. Each time, a small group of companies rode the wave and dragged the rest of the index up with them. AI is this generation's version of that story. Fifty years from now people will wonder what all the anxiety was about, the same way we laugh at 1998 headlines calling the internet a passing trend. Although to be fair, the internet never threatened to replace the people writing those headlines.
One more thought, because it matters. Even Treasuries aren't hiding from this. If AI delivers a genuine productivity boom, the economy produces more without creating as much inflation. That keeps rates low and bonds happy. But if AI displaces faster than the economy adjusts, deficits balloon and yields rise. The "risk-free" rate has quite a bit of AI risk in it too, which is something bond investors tend not to think about.
Stocks aren't expensive or cheap on their own. They're priced against bonds, and right now both sides of that equation come down to the same question: does AI grow the pie, and by how much? That is the bet. Only the name of the technology changes.