Welcome to The Week from Prof G Media, where we break down what mattered and what it all means. I'm George Hahn and it's Friday, August 14th. Today, Nvidia's half-trillion-dollar bet on AI and what a weak jobs report tells us about the economy. Then, Scott on America's declining power at home and abroad. And finally, why we're surrounded by sex but having less of it. Let's get into it.
This week, the AI boom began to look less like a software revolution and more like a heavily financed construction project. Big Tech is spending at an industrial scale. Nvidia is helping finance its own customers and no one can say exactly where the returns will come from. Last Friday, Prof G Markets ran its quarterly review with Aswath Damodaran, the NYU valuation professor Ed calls the dean of valuation. His point? That investors are now holding a very different kind of company than the one they bought 5 years ago.
These companies 5 years ago, if you asked me what their invested capital was, I wouldn't even have cared because you knew that they could generate revenues and operating income with very little additional invested capital. I'll tell you about acquisitions. Even with R&D considered, these companies generated returns of 70, 80, 90% invested capital. The only survivor from that group is Apple, which still continues to deliver that kind of return and analysts are not happy with it because it's not investing. The other companies now are the equivalent of manufacturing companies. They're building huge capacity for whatever AI products and services, and like all manufacturing companies historically, they're now going to be judged on whether they can deliver the earnings on this investment. Something they've never had to do historically. So, measures like return on invested capital that used to be not that useful with tech companies now come into play. Questions are: are you earning more than your cost of capital? A laughable question 5 years ago with these companies now becomes a relevant question.
On Monday, Scott and Ed laid out the bull and bear cases side by side. The S&P had just hit fresh record highs, but Scott pointed to a problem. The current level of AI spending would require two and a half trillion dollars in new revenue, more than what all of big tech generates today.
The problem I think we're facing and I think what people have to be cognizant of is the market is now six stocks. And Alphabet and Amazon alone drove nearly half of S&P earnings growth. It feels like the metrics are overstating the positive case, if you will. And I like to me, I think the data around the fact that things are overvalued and or the bear case, if you will, the one piece of data that I just think is very hard to argue with is that if so much of the market is concentrated on the well-being of the prospects of AI and the investments in these hyperscalers, we need if you look at the CapEx to get a reasonable rate of return on the current CapEx by the hyperscalers, we're going to need to recognize or create two and a half trillion dollars in incremental revenue from AI. Right now, the cumulative revenue being produced by AI is 150 billion. So, we need it to 15x just to get to a sustainable or justifiable ROI on the current CapEx. I think that is unlikely. And by the way, that two and a half trillion dollar number, just to put it in context, is greater than all of the revenue of Big Tech right now.
Then, Nvidia announced it had lined up half a trillion dollars from six major asset managers to fund the AI buildout. The money goes to Nvidia's customers, so they can buy more Nvidia chips. Jensen Huang called AI chips an investable asset class. Larry Fink called it the next future for financial engineering. Nvidia stock fell nearly 4%. On Wednesday, Ed asked Jay Goldberg of Seaport Global what the deal actually is.
When companies provide financing for their customers, there's a fine line between enabling demand and creating demand. And Nvidia has been walking that line ever closer for a couple years now. I think it's perfectly reasonable to provide some form of working capital loan. You lend your, you know, you give them favorable terms. You let the customer pay 6 months, 9 months. That's a form of financing. Or you loan them a little bit of money. It's common practice in a lot of, you know, CapEx intensive industries. This one I think catches everyone's eye because it's so big, right? Half a trillion dollars, like you said. It's just not entirely clear if you are coming from the point of view that AI is a bubble and nobody can articulate what the ROI on that investment is or what even the use case is or the business model. If you come from that viewpoint, this looks very much like Nvidia's creating demand that might not exist otherwise. Now, Nvidia's take is, 'No, no, we're not creating that demand. We see a mispriced asset class, if you will. The lenders don't know how to lend to this. They're missing the point. They're missing the opportunity. They don't understand it. We will step in and help bridge that gap.' I get the logic, but it's such a big number and there's so many questions around it. It's hard to see this as something that's not Nvidia giving money to customers so they can buy from Nvidia, left hand paying the right hand.
Wall Street is already pricing AI as an economic transformation. The labor market tells a more complicated story. That transformation may not have arrived yet, but worker insecurity certainly has. The July jobs report landed weaker than anyone expected. The economy lost 23,000 jobs against expectations of 80,000 gained. Unemployment fell to 4.1%, its lowest in 2 years, but only because fewer people are looking. Labor force participation dropped to 61.4%, the lowest since February 2021. May and June were both revised down. Labor economist Catherine Ann Edwards told Ed there was nothing good to pull out of the report. But she pushed back on the idea that AI is already driving the numbers.
One thing that I bring up when we talk about AI is if you were to go back and look at the average productivity of the US worker over the past 90 years, you will not see a single technological innovation. You won't see air conditioning, you won't see cell phones or computers or the internet or Microsoft Office. You don't see it because it takes a long time for productivity to be absorbed into the workplace and used by employers. That doesn't mean there's not an employer out there that is using AI, but using it effectively. Actually adapting it to the degree where you are changing how you demand your workers as opposed to just exposing them to some new technology and seeing what happens. Those are very different processes, right? Like just because your boss said like, 'Here's Claude, try it out.' There's a gap between here's Claude, try it out and we're going to lay off 200 people. Like, there's a very big step there, and most of the time it's not a jump. It's just you're kind of like crawling towards this new type of productivity. I think my kind of, you know, trying to pour a bucket of cold water on AI and its effect in the workplace is not to say that AI is not an incredible tool that will change the future, but that it's in some ways a distraction from the problem that we refuse to see, which is that the US has a very cruel policy when it comes to supporting the unemployed.
So, all of this uncertainty of the things we know and the things we don't, if you were worried about US workers, you wouldn't be looking at the bottom line of companies, you know, earnings portfolios, you would be looking at our unemployment system and demanding a better one.
