The AI Bubble: Power, Hype, and the Machinery Behind It

Every few decades, a new technology convinces us that this is the one that will change everything. The internet did it. Crypto did it. Now, it’s AI’s turn.

What makes this moment different is not the excitement, but the scale of it. The AI boom isn’t just about clever models or futuristic chatbots. It’s about who controls the infrastructure that makes intelligence possible and who pays for it.

The Money Machine

The money flowing into AI right now is unlike anything we’ve seen in tech since 2000. Nvidia alone is worth more than most countries’ stock markets. Microsoft, Google, and Amazon are spending hundreds of billions on chips, data centers, and electricity. And yet, the economic return is still tiny compared to the cost.

Most people using “AI tools” aren’t paying for them. The products are flashy, but the revenue is a rounding error next to the infrastructure bills. It feels less like a gold rush and more like everyone’s buying shovels because they saw their neighbor do it. The infrastructure is real, but the payoff isn’t, at least not yet.

A New Kind of Monopoly

The AI economy isn’t open. It’s concentrated. Two companies, Nvidia and AMD, control most of the world’s compute. Three, Microsoft, Google, and Amazon, control most of the cloud. If you’re building something that relies on large-scale training, you rent from them. There’s no real alternative.

Even “open-source” AI models depend on the same hardware and cloud pipelines. It’s a trickle-down system: the closer you are to the chips, the more power you hold. Everyone else builds on borrowed ground. That concentration shapes innovation. It decides who gets to experiment, who can afford to train, and who just gets priced out. The barrier to entry isn’t talent anymore, it’s capital and compute.

When Big Tech Funds Big Tech

The current boom is not a classic startup story. It’s an internal money loop between a few giants. Microsoft invests in OpenAI. OpenAI trains on Azure. Google builds Gemini, which lives on Google Cloud. Amazon backs Anthropic, which runs on AWS.

Capital circulates between the same players. Each one props up the other’s revenue lines. It’s a closed ecosystem that creates the illusion of competition when what we’re really seeing is coordination.

The risk isn’t that the bubble pops overnight. It’s that it never opens up. If everything flows through the same few pipelines, innovation slows. The next breakthrough will need their permission, or their hardware.

The Psychology of Inevitability

The hype isn’t just financial. It’s psychological. Every company now feels the pressure to “do something with AI.” It's a career risk not to. Job titles, roadmaps, and marketing copy all bend to fit the new storyline.

People don’t want to miss the next big thing, even if they don’t fully understand what it is. That’s how bubbles form – not through deceit, but through collective faith. The belief that if everyone else is in, it must be real.

It’s easy to forget that most revolutions look obvious only in hindsight. The hard part is telling which ones are genuine before the dust settles.

The Quiet Cost

Behind the optimism is a growing physical and economic cost. The new data centers need water and power at a scale cities can’t always support. Chip supply chains are stretched. The price of compute is rising faster than the productivity it promises.

For now, the excitement masks these constraints. But they’ll surface. At some point, the maths will have to make sense  and that’s when the narrative usually shifts from infinite to finite.

Where This Leaves Us

So, is this a bubble? Yes – but not the kind that vanishes overnight. It’s one that reshapes the landscape while it grows. The money, the compute, and the concentration are all real. So is the potential for something durable.

The story of AI isn’t just about smart machines. It’s about power, capital, and who gets to define progress. The infrastructure being built today will outlast the hype but it will also decide who benefits from it.

When the noise settles, we’ll see that the real value wasn’t in the models. It was in the machinery beneath them.

– Nankee Hari, Equanimity Investments