Jack Howe (7:08)
You don't sound like you believe me on the quickly part. Let's, let's see what happens. The first factor has to do with free cash flow. I'll call it follow the cash. Two years ago, there were nine Wall street analysts who had ventured estimates of the free cash flow that Amazon would generate by this year, 2026. And those estimates ranged from $76 billion to $126 billion. And the average was $105 billion. I know big numbers like this can all kind of blend together sometimes, but that is a staggering amount. It has been reached only by two companies ever. One of them is Apple. And the Other is the oil monopoly of Saudi Arabia. But if you look today, the 2026 consensus estimate for Amazon's free cash flow stands at $11 billion. And it's actually not free cash flow, it's cash burn. So how does a company go from plus 105 to negative $11 billion? It's definitely not because Amazon has fallen on hard times. It has decided to go all in, of course, on building AI computing power as quickly as possible. And it's not alone. Alphabet, Meta, Microsoft and others are spending astonishing sums. And you have heard discussions about whether all of this spending is wise. I think you can safely look at a company like Amazon, which already does this lucrative business in building computing power and then renting it out. For a company like Meta, which does not sell cloud computing, it might be more of a leap. But that uncertainty about how and when this spending will pay off is part of why all of these stocks that I just mentioned are down this year. Let's put aside the merits of the spending, though. I want to focus instead on the accounting. When companies spend money on, let's say, salaries and advertising electricity, that gets deducted from their sales right away for purposes of calculating their earnings. But big ticket purchases like new data centers and equipment, those go into a special category, and it's called capital expenditures, or CapEx for short. CapEx does not get deducted from earnings right away. It gets deducted little by little over years. The idea, what accountants are trying to do is to simulate for investors how that spending would have matched up with revenues if it had happened gradually over the projected useful life of the stuff being bought. It's called the matching principle, and it's at the heart of earnings accounting. Okay, hold on to that thought. It's difficult to overstate the scale of the current capex spree. Let's put it this way. Barclays, which is something of an AI bull, lately predicts that combined capex for hyperscale players like these will peak at a trillion dollars by 2028. A trillion dollars? For comparison, as recently as 2023, all of the companies in the S&P 500 combined had earnings of $1.9 trillion. If you look now, those earnings are shooting higher. The estimate for this year is 2.8 trillion. Remember, those earnings, they mostly don't subtract for all the spending that qualifies as capex. But here's the key. Earnings definitely do count all of the winnings for the companies that are on the receiving end of that capex spending. Nvidia, the AI chip King is it will likely turn the biggest corporate profit in history in its current fiscal year ending next January. The latest estimate I saw was $200 billion. 200 billion. Until three years ago, this company had never earned more than $10 billion in a single year. There are other data center arms dealers who earnings have suddenly multiplied. They include Micron Technologies, a memory maker. It's expected to make $66 billion this year. That's the fifth largest contribution to the S&P 500's total. And the chip designer Broadcom, $56 billion. That's the ninth largest contribution. So one question investors should ask themselves on behalf of companies receiving all this money is how sustainable the spending seems. But at the very least, I think investors ought to make sure that this money is sort of counted fairly when they're sizing up the market. Earnings don't really do that right now. If you want to do that, you should look at free cash flow. Unlike earnings, free cash flow is not a measure that companies are required to report on their financial statements, although many volunteer it in their slide decks and press releases. It's easy to calculate. You start with the actual cash that's generated from operating activities and then you fully subtract capex. So free cash flow for the S&P 500 is pegged at just under $1.9 trillion for this year. That's nearly a trillion dollars short of earnings. It's pretty common for the two measures to differ, but definitely not by this much. So remember I said that if you look at earnings in the price to earnings ratio of around 20, that the S&P 500 is around 20% more expensive than usual based on the past two decades. What if we look instead at free cash flow and the price to free cash flow ratio, that's 27.5 and it makes the S&P 538% more expensive than usual versus the same two decades. I think with earnings you're only getting half the story on all that AI spending and it's the good half. I think with free cash flow right now you're getting a fairer picture and it doesn't make the stock market look especially cheap. So that's factor one. I hope it makes sense for everyone. Basically you have these colossal tech companies in the S&P 500 spending unheard of sums more than a hundred billion dollars a piece. And they're spending that money with other companies in the S&P 500 and we're counting the money for the companies receiving it. But we're not Counting it for the companies spending it and it's making the stock market look cheaper than it really is, especially now.