Is The AI Bubble Really About To Burst?

Everyone's talking about this idea of this AI bubble, basically that AI is getting way more hype than it deserves. And I think it's important to understand that a bubble happens when prices rise faster than real results. And so in the case of AI, it would mean that investors start throwing a ton of money into AI companies, but those companies then struggle to get paying users and eventually die off. We saw something similar during the dot-com era, and investors were throwing money at any company with a website, even if it didn't have a real product or a way to make money. And eventually that caught up to them and the whole market crashed. Now there's a debate on whether these AI companies will actually be able to stay afloat? I'd argue yes. So much that I actually bet my friend George Gammon a healthy dose of dollars that the biggest AI companies such as Nvidia, OpenAI, Anthropic, Cognition will see more usage and revenue and greater valuation 5 years from now than they do today. So I want to break down this debate and why AI isn't following the same path as a traditional bubble market. There are a few major reasons people are calling AI a bubble. The first is that the numbers look big. You've got market caps looking really big, chip sales, funding rounds, and they are. I mean, Nvidia is the most valuable company in the world right now. But the difference is what we're seeing now is different than the dot-com bubble. Now I'm going to give you a real business use case here. So for my business, Single Grain, which is an ad agency, we do work for our clients. And one of the softwares that we built, it was saving our SEO managers 20 hours a month. Yesterday I was talking to one of our SEO managers and she was saying that now it's saving them 40 hours a month. I'll give you one more over here with AI, with the product that we created, ClickFlow, that I talked about earlier. It has actually created pages for us. One of those pages led to a client that is now driving us about $50,000 in monthly recurring revenue. And that is the power. That is the actual business use case that we're seeing from this stuff. Now let's talk about something that happens when something becomes easier or cheaper to use over time. People use it more. Look at electricity. When it wasn't available in the past, it was super expensive in the past. Well, what happened? The cost came down. It's more deflationary over time. More and more people started to use electricity. And you look at today, we don't have enough energy. We need more electricity, especially to power the AI that we need. We're actually talking about computing here. We're building more computing. What is computing? It's actually intelligence. And so we're all trying to buy more intelligence here and we need more energy for it. People are just able to use more intelligence. When I talk to some of my doctor friends, they say that it's way more efficient than they are as one person and they often going to AI and asking questions, and they're using AI to help diagnose things. And by the way, when you go to a doctor, what are they giving you? 15, 30 minutes of their time. They're not able to research as much as you want. And so you as a consumer, you are now partnered up with a super doctor in your pocket, and you can then partner with AI to then work with your doctor who's also working with AI. The demand for energy is actually Jevons paradox happening. Jevons paradox means AI is going to become used more and more. And if you're seeing The use cases in your personal life, and then you're seeing it work with businesses as well. It makes sense that we're going to need a lot more compute, and it's going to be used way more. And another reason why AI isn't in a bubble right now is that AI isn't just helping us do old work faster. What do you see on the streets right now if you're in Los Angeles, or if you're in China right now? What are you seeing? You're seeing self-driving cars. The driving is old work, but what about the new work that's going to happen? We built an AI agent, an agentic workflow that helps coach our people in terms of prioritizing on the right things. We couldn't do that before. Now we're able to do that and we coach in public. So these aren't going to be replacing humans because this is work that humans weren't doing in the very beginning. And by the way, speaking of old work, people forget that AI hasn't even really hit the physical world yet. I mean, sure, you see self-driving cars like I mentioned earlier, but robots, humanoid robots, I mean, Tesla's building these robots right now. You got companies like Figure that are building these robots as well. Smart homes are, are, are the early contenders as well. But once intelligence moves from just screens to real-world systems, the compute and token demand is going to explode. And that next layer is coming fast. Now look, I want to give you another story here. So I host a lot of these AI founder dinners along with one of my friends, Ryan, and we try to do these once a month. We're hearing the same stories overall. I actually, I just did one of these dinners 2 weeks ago, and one person mentioned that what used to take 30 engineers, it can now be done with the work of 2 engineers. And so that just means that the real efficiency gains are here. This is not some pie in the sky type of stuff. You see all these engineers, the reason why a company like Cursor is growing so quickly doing hundreds of millions in revenue is because engineers are using this to help them complete code faster. I think these stories should tell you that this stuff is real. And when I'm at these AI founder dinners, so many of these people are testing and they know that this is the future. They know this is the frontier and they're very excited about it. If you want to run personalized LinkedIn ads at scale and get these ads up in minutes instead of weeks, you've got to check out Karrot. So Karrot, K-A-R-R-O-T.ai, will allow you to create personalized ads on LinkedIn, personalized landing pages. We are also adding in different automations to spin up these ads quickly. We have different creative options as well that will integrate even cool things like Google Nano Banana. A lot of cool things coming down the pipeline. If you are serious about B2B marketing, you're serious about LinkedIn, you've gotta check out Carrot and we'll see you on the other side. Look, I'll share another story with you. When I get on some, some of these sales calls just to listen to what customers are looking for, the conversation has now shifted to replacing headcount. So I was talking to a company recently and the global CMO just told me on the call, he's like, You know, look, if the conversation shifts to for the price of one headcount, we're doing what 4 people can do or what 8 people can do. That becomes a conversation that's much more interesting to us. And so when you look at the labor budgets around the world, that's $13 trillion a year. When you look at marketing budgets, it's just a small piece of that, right? But when we talk about labor budgets, that's a completely different conversation because payroll is oftentimes the most expensive thing on a business's profit and loss statement. Now, look, my friend George has an argument that this all kind of looks like Enron, the round tripping of revenue. You know, companies that are investing in other— you got, you got You got OpenAI that is putting money into Oracle, and then you have Nvidia that's also putting money into into OpenAI as well. And then you know OpenAI is also investing in Nvidia's chips, and then it's it's all being round-tripped. But here's where that breaks down: real usage is driving all of it. Tokens are being burned, compute is being consumed, and businesses are paying for output that didn't exist before, and they're getting real-world results from it. It's driving revenue, and it isn't necessarily increasing headcount for them, but it's driving growth for these companies. And that's not a closed loop. There's real market demand. When you talk about Enron, you talk about overbuilding infrastructure in the past, and a lot of those companies have gone bankrupt, but we're using that infrastructure today, right? Now, when you look at Enron, that was an accounting scandal, right? They reported on growth that they didn't actually have, right? And so that, that's completely different. But when you think about intelligence, we've always needed more intelligence. When you think about businesses, businesses are what? They used to always hire more people. Well, people are what? Fundamentally, we're looking to hire more intelligent. I completely believe that the world's going to be completely different. And I'm excited to come back to this video in 20 years to see how things have changed. So we've got another argument from skeptics. And my, my buddy George is one of those skeptics. And he's like, how are you going to grow into these valuations with $20 subscriptions? Like the math just doesn't work. And I would just say that that's not how tech works, right? Especially if you are a hypergrowth company and you think about Google back in the day or Facebook back in the day, now Meta, these are land grabs. And what I mean by that is you're looking to acquire as many users as possible, and sometimes you don't look to monetize until later. Uber's a really good example of this, where people are like, Uber is never going to make money, right? But you focus on acquiring the users first, you focus on driving real-world adoption, you look at the engagement rates, and then you figure out based on how people are using your product how you can then monetize, right? And they're already monetizing on $20 subscriptions, they're monetizing on $200 a month subscriptions. But that's not all. You think about ads, OpenAI, ChatGPT, they have 800 million users right now. That's going to continue to go up. It is one of the fastest growing companies ever. And so when you look at this, they will figure out how to monetize in the long term. Sam Altman, he has funded some of the best companies in the world when it comes to talking about feedback loops, talking about how to grow hypergrowth companies. He knows what he's doing. So you want to build a user base first, then you can worry about enterprise API ads. And OpenAI is already following that pattern. When you look at Pulse, And what Pulse does is it'll look at all the things that you talked about and it'll make a newsfeed for you based on what's interesting to you. And then eventually, what do you think is going to happen with that? They're going to run ads there. Here's what I want to reemphasize again. Beyond what we just talked about here, we're just getting started. Every week, more people are starting to use AI to write, to code, design, analyze, and build. When I visit my dad or I'm talking to family members, they're not really touching it. Sure, they've heard of ChatGPT. They might have used it a little bit. I use it for everything, but I know I'm just scratching the surface right now, and I know the world is just scratching the surface. And as we get more compute online and as these models continue to progress, based on Jevons paradox, the token usage is going to go way higher than what it's at today. And even if AI progress stopped today, it would still be phenomenal. But I'm telling you, this is just the very beginning. So here's the takeaway for you. A bubble pops when there's nothing real underneath it. What we're seeing in AI— real usage, measurable output, cost savings, adoption across industries— That substance. And yes, prices are high, but the fundamentals are growing. That's not a bubble; that's an early stage boom. It is irrefutable. So look, if you watch this all the way through, I'm curious to get your take. Drop your thoughts in the comments, and I will see you in this next video over here on AI and marketing.