Will AI be a labor productivity booster or worry in 2026?
00:00 Speaker A
Oh Russia, one of the topics that came up at the press conference yesterday was artificial intelligence. As an economist, I’m curious about this wonderful trend and theme when you just think about it with your kaleidoscope. Russ, what do you think about the impact of this on GDP and the labor market?
00:23 Russian
Yes. So the labor market has stalled, and the question is: Is Artificial Intelligence part of that pause? Are companies now getting much more productivity from the employees they have? They don’t need to make any more new hires, or it could be because they’re struggling with tariffs or seeing a drop in overall demand. I think the overall picture largely reflects a situation where small businesses are cutting jobs or halting hiring, and that suggests to me that the problem is more on the tariff side of the equation. Yes, there will be job losses due to AI, but I don’t think they’re quite at the point yet where they will be the driving force behind the labor market. I think it will be a much longer period of time. uh and referencing the Fed’s dot plot, I think it would be a pretty strong productivity generator and would allow the economy to grow at a better than normal sustainable rate, keeping unemployment low and inflation in check.
01:50 Speaker A
David, let’s stick to the AI theme because what you’re saying here is that we’re definitely in an AI bubble. Let’s stop here, David. What are you doing, I guess two questions. What do you mean by bubble, David? Describe this for us. So what are the metrics, David, what are the indicators that tell you that yes, we are together?
02:19 David
oracle Right? So you get overexcited, you get overdebt, you build overcapacity, and then there’s a period of time where Mitch is out of sync with how quickly all that revenue is coming in, right? If we need to produce $3 to $4 trillion worth of stuff between now and 2030, then to monetize that with a 10% ROI, we need to generate an extra trillion dollars a year in revenue through hyperscalers, and that’s not going to happen. Here’s an interesting statistic about Mag 7 this year. Only two of the Mag 7s outperform the 493 index. and the Russell 2000 fulfills both in Mag 72. So I think capital is already leaking from the AI business. and if you look at CDS spreads and things related to Oracle, they have negative cash flow. They will spend 75% of their revenue on building data centers for open AI. I mean, they might have an Icarus complex to turn on the AI. I really don’t know. So I think there are other places to go, and as the dot chart suggests, if we’re going to get two and a half percent growth in GDP next year, you don’t need to go after AI. You need to be in finance, you need to be in materials, you need to be in industrials in the S&P 500. This is not a big wait, but within the 600 and 400 S&P indices, these are very big waits, 2x waits, and that’s why these indices are rising. There’s no reason to go after these overleveraged names. I hope they build a bunch more tailgates and next year it’s up 25% or 30% as they say, because that means you’re going to sell more cement, you’re going to sell more steel, you’re going to sell more copper. There are so many ways to leverage this as it moves beyond the initial hyperscalers, and you’re already seeing that in the indices this year. and then one last statistic and I’ll stop. There is $60 billion or a trillion dollars in the S&P 500. There is approximately $5 trillion in the S&P 400 and S&P 600. All you have to do is do some rotation and get the kind of spreads you see today.
