Are We Living Through the AI Bubble Again?

tobylucich
6 days ago
3 min read

If you hang around business leaders long enough, you start hearing the same conversation on repeat. Everyone says they’re “all in on AI,” but if you listen closely, they aren’t talking about the same thing at all. It’s like two people arguing about a train… except one means the railroad, and the other means CrossFit.

That’s what this week’s conversation reminded me of — the moment in history when the world went all-in on infrastructure, only to realize the brakes had been quietly removed.

…

The Telephone Card That Cost 10 Cents a Minute

Back in college, calling home wasn’t free. It wasn’t even cheap. You dialed an 800 number, punched in a long-distance calling card, and literally watched your money disappear at ten cents a minute.

Telecom was a cash cow. The long-distance rackets were printing money. Then came the dot-com boom, and telecom companies raced to lay thousands of miles of fiber to keep up with growing demand. Everyone believed they were building the backbone of the future.

And then—almost overnight—the industry cratered.

All that fiber hit the ground, the demand didn’t grow fast enough, and prices collapsed from $10,000 a month for high-capacity lines to under $1,000. Infrastructure companies went bankrupt by the hundreds. The technology was real, the future was real, but the timing was wrong.

And ironically, the real innovation wasn’t even more fiber — it was better hardware and software that squeezed more performance out of every existing strand.

…

So Why Are Investors Like Michael Burry Shorting AI Stocks?

In the last few months, the same names from the 2008 housing collapse have reappeared. Michael Burry. Peter Thiel. Insider executives quietly moving to cash. Meta’s own chief AI scientist contemplating an exit.

These are not “doom and gloom” people. These are data-driven, pattern-recognizing people. The moment the crowd gets loud, they start listening for the quiet sounds underneath.

And the quiet sound underneath AI right now?

It’s the same one telecom heard in 2000:The infrastructure is scaling faster than the viable business models.

AI takes enormous amounts of compute, water, energy, and data-center footprint. Someone has to pay for that. And while CEOs are pouring billions into AI because it feels like the future, the returns aren’t showing up yet.

Sure, people are using ChatGPT to make gluten-free recipes and Christmas lists — but that’s not a business model. Not one that justifies trillion-dollar chip valuations.

…

The Process Debt Hidden Inside the AI Hype Cycle

When a company invests in AI, what they want is upside: new products, new revenue, new markets. What they actually get (at least right now) is cost: infrastructure, compute bills, consultants, and a handful of “innovation projects” that don’t go anywhere.

That’s not AI’s fault. That’s a process problem.

We’ve seen the pattern before:

The first movers innovate the technology.
The second movers figure out how to monetize it.
The late movers scale it sustainably.

AOL made the internet accessible — Google made it profitable.Netscape made browsing cool — Microsoft made it a platform.Blockbuster built distribution — Netflix reinvented delivery.

If the pattern holds, GPT-3.5 or GPT-4.1 might end up powering more value than GPT-5.

Companies are so busy chasing “the next big model” that they’re missing the lesson hiding in plain sight:Innovation is expensive. Useful innovation is affordable.

And in the middle, there’s always a crash where someone pays the piper.

…

So What Happens Next?

Here’s my bet: Consumers will keep winning. AI will get faster, cheaper, and more integrated into everyday work. But a whole lot of companies — especially those betting on limitless growth in compute — are going to feel the sting that telecom felt in 2000.

Not because AI isn’t real.But because process always wins.

And right now, the process side of AI — how it makes money, how it creates value, how it reduces work instead of adding more — is still totally undeveloped.

That’s the real conversation CEOs and employees think they’re having, but aren’t.

…