I've been thinking about the AI bubble for months now.
Not whether it exists. It definitely exists. But what happens when it pops.
This is Part 1 of a series where I'm going to break down what's actually happening in the AI market, what history tells us about bubbles, and what it means for people like us who build with these tools every day.
Let's start with the basics. What even is a bubble?
Look At Streaming Services Right Now
Think about how many streaming services you have.
Netflix. HBO Max. Disney+. Hulu. Peacock. Paramount+. Apple TV+. Amazon Prime. Discovery+.
The list goes on.
Every single one of them has been losing money. For years. They're all exposed to subscription fatigue from people who can't afford eight different services at $15 each.
Nobody can sustain it. Everyone knows it.
But here's what most people miss. They're not trying to be profitable right now. They're trying to outlast each other.
The War Of Attrition Strategy
This is how bubbles actually work.
You burn cash. You lose money. You keep operating even though the math doesn't make sense.
Why? Because when your competitors run out of money first, you absorb them. Their content. Their subscribers. Their market share.
Suddenly you're not splitting the pie ten ways anymore. You're taking most of it.
And when you own most of the pie, you set the prices. You own the market.
That's a bubble. Companies burning cash in a war of attrition, hoping to be the last one standing when the music stops.
This Isn't Theoretical Anymore
Netflix just announced they're acquiring Warner Bros for $82.7 billion.
HBO. Harry Potter. Game of Thrones. DC. Friends. All those heavy hitters.
Warner Bros couldn't sustain the losses any longer. Netflix could. Now Netflix absorbs them.
The streaming bubble is consolidating right now. In real time. This is what the end of a bubble looks like.
Companies that couldn't outlast the cash burn are getting acquired by companies that could.
AI Is Following The Exact Same Pattern
Look at the AI market right now.
OpenAI. Anthropic. Google. Meta. Mistral. Cohere. Dozens of smaller players.
All of them burning massive amounts of cash. All of them losing money on every API call. All of them subsidizing usage hoping to outlast the competition.
Sound familiar?
The same pattern is forming. Too many players. Unsustainable economics. Everyone waiting to see who runs out of money first.
Why This Matters For People Building With AI
I use these tools every day. Claude for automation projects. GPT for client work. Local models for specific use cases.
Right now, they're incredibly cheap. Almost too cheap.
ChatGPT Plus is $20 a month. Claude Pro is $20 a month. API costs are subsidized to the point where they don't reflect actual operational costs.
That's bubble pricing. They're buying market share with investor money.
But that doesn't last forever. At some point, the math has to work. At some point, prices go up or companies go under.
If you're building business systems that depend on these APIs, you need to understand what's coming.
The Pattern Always Plays Out The Same Way
Bubble forms. Money floods in. Too many competitors. Unsustainable economics.
Bubble pops. Most companies fail. The survivors consolidate the market.
Then prices go up. Because the survivors don't have competition anymore.
We've seen this with streaming. We saw it with the dot-com boom. We're watching it happen with AI right now.
The question isn't whether the AI bubble will pop. It's when, and what happens after.
That's what this series is about. Understanding what's actually happening so you can make informed decisions about how you build with these tools.
Next up in Part 2: What the dot-com bubble actually looked like, and what it tells us about what's coming.