The Biggest Risk After Nvidia Earnings

**Dave** (0:04)
Last week, all anyone could talk about was how we are in an AI bubble. How there's way too much hype, way too much spending, no real returns. The market pulled back about 3%, big tech got hit the hardest. Nvidia was down 7% since November 1st. All gloom and doom. Until Nvidia enters the chat last night, and all of a sudden, nobody's talking about bubbles anymore. Nvidia beat earnings estimates all around $57 billion in revenue. They're expecting $67 billion next quarter. They said Blackwell demand is off the charts. That's a direct quote. They said cloud GPUs are basically sold out everywhere. No surprise. Stock jumped 5% after hours, which at this scale, that's $200 billion in market cap in a matter of minutes. The ripple effect was widespread, small caps were up, crypto was up, biotech was up, the tone completely flipped. People on X were trolling the AI bubble, laughing at Michael Burry's latest big short, pointing to the hyperscalar backlogs that are now reportedly over $500 billion. And there's speculation that Trump could reopen China, which would mean another $5 billion a month in demand for Nvidia. Basically, the narrative went from AI bubble to AI tsunami overnight. But there is a real risk out there, and it's not the obvious stuff that everybody always talks about. Today on Dumb Money, we're going to reveal the biggest risk after Nvidia's earnings and the trade setup we're about to play. This is Dumb Money Live.
Hey there, Dave here, along with Chris and Jordan. We are Dumb Money. Welcome to Dumb Money Live. Quick reminder, smash the old like button for the almighty algorithm. It really does help people find the show. It helps YouTube know that we're not wasting our time here this morning. Chris, Jordan, Nvidia didn't just beat, they kind of reset the narrative briefly. But as we're watching in real time, that's all falling apart. We were watching this reaction in real time last night. What were your first takes on Nvidia's earnings and what's happening right now?

**Chris** (2:05)
Well, listen, Nvidia earnings were exactly what I anticipated them to be, which were unbelievably great. We knew this, I feel like everybody knew this, but what's the takeaway is that there was so much concern going into earnings. There's so much anxiety and so much fear and so much doubt on this AI cycle, that whatever happens, someone is going to find a hole in it, right?

**Dave** (2:43)
And fear and doubt, the fear doubt index was like pegging all the way on the fear side, or the, what is it, the fear and greed index was like all the way in fear, like further than we've ever seen the needle go.

**Chris** (2:57)
Yeah, I think we learned a lot over the last week, and what we learned was the psyche of investors right now, both institutional and retail. It is shaky, and it doesn't really matter what Nvidia earnings are. It really doesn't matter how great Gemini 3 is. It really doesn't matter the degree to which people like us can clearly see the biggest thing that's ever happened in maybe all of, most of humanity, like recent humanity, right? And it's just so obvious to us, none of that matters, because ultimately you have investors that have an enormous amount of doubt around what AI is, what it can be, the degree to which it will be monetized, the timelines, the economics, the circular financing. There's, I've never seen so much.

**Jordan** (4:00)
The finance, I think the valuations, too. I mean, we saw Michael Berry shut down his hedge fund because he just doesn't, he can't get over the valuation mismatch that he sees.

**Chris** (4:10)
Dude, he's couldn't get over any, here's the thing, dude. I don't even know where to start with Bert.

**Dave** (4:17)
Yeah.

**Chris** (4:17)
Like he's been saying-

**Jordan** (4:18)
I heard a rumor, I know you can talk in a second, but I want to say one thing, and the one thing that he called out, or I think that he called out, I didn't actually read his letter, I'm sure this is second-hand, but that companies like CoreWeave, that their depreciation model, he thought was way off, that they're not depreciating these assets, these GPUs nearly fast enough for what their useful life should actually be.

**Dave** (4:43)
He called out that they had a depreciation schedule, and over time, they've been increasing that depreciation.

**Jordan** (4:50)
They have been increasing it, right? But he saw like a 5% on the first year, depreciation schedule for some of these, and he was like, whoa, that is not near enough.

**Chris** (4:58)
Okay, but what we're finding out is that the demand for inference is higher than anybody anticipated it would be. The supply glut on compute is deeper than we ever thought it would be. Demand for compute is higher than anyone ever anticipated that it would be. And because of that, it still makes economic sense to utilize these chips in years four, five, six, even if you're utilizing them for things that you ordinarily didn't really have on the books, like you weren't using them for those use cases, right? And so-