Maybe we're not all doomed: the AI bubble is a bottleneck story

Credit first. This post is mostly someone else's thinking. It is built on the video "Why Everyone Is Wrong About the AI Bubble" by the Maxinomics channel, published December 22, 2025. This is mostly their thinking: they did the research and made the argument. I am retelling it in my own words and adding what it means for a small business. Every number below is theirs, not mine, and I have not checked any of them against a primary source. Watch the video; it is better than this post.

The question everyone is asking

Is AI a bubble? Is the whole thing about to pop the way the dot-com boom did in 2000 and 2001? If you run a small business, you may be asking a more practical version: should I wait and see?

I am not going to tell you whether stock prices are right. I do not know. What I found useful in the video is how it explains why the dot-com crash happened, because the reason is not what most people remember.

The dot-com crash was a bottleneck

Most of us remember silly websites. The video says the real damage was somewhere else: in the telecom companies that laid enormous amounts of fiber optic cable, far more than anyone needed yet. It cites around $500 billion spent by 2001, with less than a tenth of it carrying any traffic. The rest sat dark.

The cable was fine. The problem was the last mile, the stretch of old copper phone line between the fiber and your house. Web pages took forever to load. Going online meant nobody could call you. Millions of homes could not use what had been built, so demand never showed up to fill it. Money poured into one part of the chain while the weakest link stayed exactly as it was.

That is a bottleneck story. The whole system could only move as fast as its slowest piece.

AI's last mile is power

Here is where the argument turns hopeful. The video says AI is different in one important way: the capacity being built is not sitting idle. The chips are running around the clock, so there is no dark Nvidia chip waiting for a customer. People are using the products, and cheaper AI has tended to produce more use, not less.

So where is the weak link this time? Electricity. The video says the power the announced data centers need would nearly use up the spare capacity the US grid keeps in reserve. In its words, "we need more data centers is really code for we need more electricity."

I read the takeaway as a yellow flag, not a red one. If the power gets solved, the bubble worry looks overblown. If it does not, the whole thing stalls, the same way the internet stalled on copper wire. Nobody has to be a fool for that to happen. One link in the chain is just too slow.

I like this framing because it replaces "doom or hype" with a question you can actually ask: where does it get stuck?

Now bring it home to your business

This is the part I know about. When a small business tries AI and nothing changes, the technology is almost never the reason. The tools work. The model is fine. Something else is holding the work up:

A follow-up that only happens when one person remembers it.
The same customer details typed into three different places.
An answer that lives in one employee's head.

Better AI does nothing for those, any more than more fiber helped a house on a copper line. You can buy the best tool on the market and still have the same slow spot in your week, because the tool was never the slow spot.

That is why I do not start with technology. I start by asking where your work gets stuck.

How to find yours

You do not need a consultant for the first pass. Try this:

Think about the last time something took longer than it should have. A quote, a callback, a handoff.
Ask what it was waiting on. Usually it is a person, a re-typed field or a missing piece of information.
Ask whether the same thing happens every week. If it repeats, it is a bottleneck. If it happened once, it is just a Tuesday.

Ten minutes with a notepad is enough to start.

What I do with it

Once we find the bottleneck, I build the fix so it stays fixed. That means a fixed quote before work starts, you own what gets built, and it keeps working when the tools around it change. Then we look for the next slow spot. That is the whole method, and it is a deliberately small one.

So, are we doomed?

I do not know what will happen to AI stock prices, and neither does anyone selling you a prediction. What the video gave me is a calmer way to think about it. Big systems fail at their weakest link, and weak links can be found and fixed. That was true of the internet in 2001, it may be true of AI's power supply now, and it is definitely true of your Monday morning.

Show me yours

Bring me the bottleneck in your business and we will automate it for good. On the free call you describe your week, I look for the repeat work, and you leave with one place to start. No obligation.

Show me yours. Text (512) 337-9530, email josh@joshdlange.com, or book the free AI assessment.

Source: Maxinomics, "Why Everyone Is Wrong About the AI Bubble", YouTube, December 22, 2025. All statistics are the creator's claims and were not independently verified. Checked October 14, 2026.