If you’ve paid even a little attention to the news, you’ve heard the buzz. Artificial Intelligence isn’t just a futuristic movie plot anymore; it’s here, and it’s fuelling a market frenzy that’s hard to ignore. Companies like Nvidia, which makes the essential chips for AI, have seen their value explode, recently surpassing a $5 trillion market cap, making it more valuable than the entire economy of almost every country on Earth. It all feels a bit familiar, doesn’t it?
It’s impossible to see this without getting flashbacks to 1999, right before the dot-com bubble burst, wiping out trillions in investments. So, what’s the deal? Is this another case of speculative hype destined for a painful crash, or is this the real, foundational build-out of our future? The truth is, it’s not one or the other. It’s both.
Right now, we are witnessing two things at once: a classic, high-risk financial bubble sitting on top of a very real, very significant technological revolution. Let’s break down what’s really going on.
“This Feels Like 1999”: The Case for a Dotcom Bubble
If you’re skeptical, you have every right to be. The financial warning signs are flashing bright red, and they point to a market that has become detached from reality.
The Numbers Just Don’t Make Sense
In a normal market, a company’s value is tied to its profits. In the 2025 AI market, not so much.
- Insane Valuations: Take Palantir, an AI-focused company. It’s been trading at 700 times its earnings. In a normal growth market, a 25-40x multiple is considered high.
- Massive Losses: Even the companies at the very center of the boom are burning cash at a stunning rate. In the first half of 2025, OpenAI (the creators of ChatGPT) reportedly posted a staggering $13.5 billion loss on just $4.3 billion in revenue.
- Unprofitable Customers: Here’s the real kicker. For many top AI products, like Microsoft’s GitHub Copilot, their best and most active customers are actually unprofitable. The computing cost to answer their complex questions is more than the monthly subscription fee they pay. Think about that: the more people use the product, the more money the company loses. That is the definition of an unsustainable business model.
The “Trillion-Dollar Loop”
So, if these companies are losing so much money, where is all the “revenue” coming from? Much of it comes from what analysts call a “circular flow of investments.”
It works like this:
- A tech giant like Nvidia or Microsoft invests hundreds of billions of dollars into an AI company like OpenAI.
- OpenAI then uses that exact same money to sign massive, multi-billion dollar contracts to buy AI chips (from Nvidia) and cloud computing services (from Microsoft).
It’s a self-feeding hype machine. It looks like revenue on the books, but it’s really just the industry’s biggest players passing the same money back and forth in a closed loop. The real question is: what happens when they stop?
The Tech Is Hitting a Wall
For all the talk of “Artificial General Intelligence” (AGI) being just around the corner, the technology itself is showing signs of diminishing returns.
This is what some call the “Peak GenAI” thesis. The much-hyped release of GPT-5 was considered a “dud” by many users, showing only minor improvements. Worse, the core “hallucination” problem, where the AI confidently makes things up, is still rampant. This isn’t a solid foundation for a multi-trillion dollar industry.
“This is Electrification”: The Case for a Real Boom
Now, before you sell all your stocks, here is the other side of the story and it’s just as compelling. The argument here is that this isn’t 1999. This is 1899. We’re not building frivolous websites; we’re building the new electrical grid.
The “Adults” Are in Charge
The dot-com bubble was driven by thousands of brand-new, unprofitable startups with no business model, all funded by speculative venture capital.
Today’s AI boom is being led by the most established, profitable, and powerful companies on the planet: Microsoft, Google, Amazon, and Meta. They are funding this revolution from their own massive cash flows, not just risky outside capital. As Goldman Sachs research has pointed out, these high valuations are, at least in part, based on strong fundamentals and real earnings from their core businesses.
They’re Building “Digital Railroads,” Not “Fields of Dreams”
This is the most important counter-argument. The trillions being spent on AI infrastructure (mostly data centers) may seem crazy, but it’s fundamentally different from the 1990s fiber-optic bubble.
Back then, companies speculatively laid thousands of miles of fiber-optic cable hoping customers would one day show up. Most didn’t, and the companies went bankrupt.
Today’s data centers are not speculative. As investment firm KKR points out, a massive data center only gets built after a giant like Amazon, Google, or Microsoft signs a long-term, legally-binding contract to rent it. The demand is already locked in.
Furthermore, unlike speculative fiber, you can’t have an infinite glut of data centers. Their construction is limited by very real-world constraints: land, permits, and most importantly, access to a staggering amount of electrical power.5 This physical barrier prevents the kind of over-supply that destroyed the dot-com market.
The “GenAI Divide”: Why Is AI Failing for So Many?
This brings us to the central paradox: how can AI be both a revolution and a failure? The answer is what’s being called the “GenAI Divide.” The reality on the ground is that while 88% of companies are using AI, the vast majority are failing to get any real value from it. Shocking reports show that 95% of all AI pilot projects “crash and burn,” delivering zero measurable return on investment.
But…
The 5% of companies that do successfully cross that divide are seeing transformative, world-changing results. For every $1 these successful adopters spend on GenAI, they are seeing an average return of $3.71.
We’re already seeing it in the real world:
- Mercedes-Benz is building cars that can hold natural conversations with their drivers.
- Mercari, Japan’s largest online marketplace, anticipates a(https://cloud.google.com/transform/101-real-world-generative-ai-use-cases-from-industry-leaders).
- Colgate-Palmolive is using AI to accelerate new product development.
The Verdict
So, will the AI bubble burst? Yes. The financial bubble almost certainly will. The “trillion-dollar loop” will break, companies with no path to profitability will go bankrupt, and stock prices built on hype will come crashing back to Earth.
But it won’t be an extinction event like the dot-com crash. It will be what one expert calls a “clarification”. When the dust settles, the 95% of failed, hype-driven companies will be gone. But the “digital railroads”—those contracted, power-hungry data centers will still be standing, fully paid for and ready for the next wave of real, profitable innovation.
The only thing that could truly stop the boom is something outside of finance: geopolitics. The entire AI supply chain is perilously fragile. It depends on China for 85-90% of refined rare earth minerals, Taiwan for the most advanced chips, and a single company in the Netherlands (ASML) for the one-of-a-kind machines that make those chips. A disruption in any of those three points could halt the entire revolution overnight. Barring a geopolitical disaster, our future is clear: we are in a bubble, and it will pop. But the AI revolution its funding is very, very real.
