Oil Prices & the Midterm Election Loom Over the AI Trade

Sam Vardis (0:00)

Welcome back to Market On Close. I'm Sam Vardis on the floor of the New York Stock Exchange to discuss the future of the AI trade and the impact of elevated energy prices on the market. Let's welcome in our next guest, Shawn Snyder, Economic Strategist at Potomac Fund Management. Shawn, thank you for joining us this afternoon. I mean, the difference a week can make. And last Monday, we were all freaking out about, you know, coordinated AI slowdown talks, and now we're seeing stocks absolutely ripping. There seems to be some traction on Meta's Muse AI agent, we've even had some news about Anthropic having to reportedly delay its IPO.

But pulling it all together, I'm just wondering how you're thinking about the economics of all of this.

Shawn Snyder (0:42)

When I think about the economics of the AI trade, I don't necessarily think that a slowdown is the end of the world. And the reason I say that is when you think of the tech bubble, what happened during that period is valuations just weren't justified because profits weren't keeping up.

So that's why we saw that slowdown eventually happen. In this case, I think news like we got today from Meta, not only just being able to spend money on CapEx, but able to monetize it through Muse, I think that's a positive. And if you think of it like running a marathon, eventually you have to settle into a slower pace in order to finish the race. You can't just sprint the whole time. So I think it's probably okay to take a little bit of breath and to pause a little bit and for things to slow down a little bit because that reduces the odds of an AI bubble. I also think market likes lower oil prices, which is what we're seeing today.

Sam Vardis (1:33)

Yeah. Let's just stick with the slow down talk first. We'll get on to oil in a second because I'm wondering if it's not the end of the world and it is being seen as a positive, what would the slow down look like? I mean, would it look like the vision or view that Sam Altman has or more like a Mark Zuckerberg type take on it, where the owner should be on the company as to why they actually held back Muse for several months to try to get the security and the safety right?

Shawn Snyder (2:04)

I'm not going to dive in the specifics on that. The way I'm looking at it is just through the CapEx spend. I tend to think of the CapEx spend as a lot like inflation. Investors tend to care about the pace, not the level, even though the regular citizen cares about the level. At the end of the day, you're seeing CapEx spend in AI hyperscalers growing at about 95 percent year-on-year in 2026

Next year, that slows to 40 percent. The following year, it slows to potentially 13 percent, and then by 2029, you're looking at five percent. I just think over time, you're going to go from this period of hypergrowth to normalization. I think this discussion of regulation is part of that. I think this discussion of guardrails is part of that. I think we're just in the early days of seeing a normalization of the AI space.

Sam Vardis (2:52)

Okay. Bringing it all together, there's just so many parts of this conversation. We obviously had Fed rate hike last week. They're penciling in more to come. There is a view out there and an argument to suggest that that is going to hurt parts of the economy that are already struggling and could then prompt an even bigger dependence on the AI part of the story that is helping the economy right now. If we do see a slowdown, would that not be a risk to overall growth? I mean, how are you seeing it in the sort of bigger picture right now?

Shawn Snyder (3:26)

Potentially. So if you look at those hyperscalar forecasts for AI spend, it ratchets up to about 1.6 trillion and it stays there for a while. And if we look at the tech bubble in 1990s, that spend on investment and that investment cycle was adding about 0.5 to 0.8% to GDP for about five years. And then we got to 2000 and 2001 and it reversed and it actually dragged down growth by about 1%.

Right now, AI spend is doing pretty much a similar story. It's adding about 0.6% to GDP, which is about 30% of real GDP. So eventually that probably will slow, but I don't think we're there yet. If you look at the CDS spreads of these companies, they're still quite low. And I think there is the capacity to continue to spend.