In the era of AI, what happens if the AI bubble bursts?
Some economists and people of the general population argue that we are in an AI bubble that is inevitably bound to burst.
Why the AI bubble bursting isn’t the end
If the AI bubble does burst, it probably won’t mean the end of Artificial Intelligence. It would likely trigger a classic industrial transition. AI would stop being the thing that everyone is betting on, and become something integrated into our daily lives without thinking twice, much like electricity or wi-fi.
Historically, major innovation booms have followed a well-documented boom and bust cycle. Major technological shifts tend to follow the same pattern: a revolutionary innovation attracts a flood of speculative money (the boom), which leads to overbuilding and a financial crash (the bust).
What gets talked about less is what comes after: the long stretch where that overbuilt infrastructure quietly becomes useful to everyone. I call that part the buildout.
We’ve been here before
Economists and financial historians regularly compare the current AI expansion to some defining boom and bust cycles in modern history. There are many, but the most notable comparisons, from my perspective, are British Railway Mania in the 1800s, and the dot-com bubble in the late 1990s.
British Railway Mania
The first big railway opened up in 1830. It worked well, and its owners made good money.
The Boom: Buying railways shares was easy and seemed lucrative so people rushed in to invest. Parliament also approved to build far more tracks/lines than the country really needed.
The Bust: Interest rates rose, share prices crashed, and investors were in trouble. Many lost their money, and the biggest railway boss was caught faking his profits.
The Buildout: Building carried on through the crash. The network grew, and Britain ended up with a national railway that lasted. Travel and shipping got faster and cheaper, which reshaped where people lived and how they were connected to the world.
The internet and dot com bubble
The internet opened up to the public in the early 90s. Back then, Netscape was the company behind the most popular web browser. They went public in 1995, and its shares more than doubled on its first day. Everyone wanted in!
The Boom: Investors poured money into anything with a “.com” in its name, sometimes even if they didn’t have a solid business plan or path to profitability. Telecom companies laid enormous amounts of fibre optic cable, betting that internet traffic would keep exploding.
The Bust: interest rates rose, and in March 2000 the stock market started to fall. The tech-heavy NASDAQ lost almost 80% of its value by 2002. BIg names (at the time) like Pets.com disappeared.
The Buildout: The cable was already in the ground, and a lot of it had never been used. After the crash it was bought up for cheap, which pushed down the cost of getting online. The companies that survived, like Amazon and Google, grew on top of that cheap infrastructure, and it paved the way for streaming, social media, and many other components of the online world which we now don’t even think much about.
The pattern behind the bubble
Looking back at a couple centuries of industrial transformations, you can find quite a few cycles like this where a consistent theme seems to hold: innovation being real and beneficial to humanity doesn’t prevent the bubble from forming and bursting.
Railways did change transportation. The Internet did change humanity. The technology was never the problem, the timeline was.
In each cycle, the market looked at a real revolution and assumed the explosion would just… keep exploding. Decades of growth got crammed into a couple of years' worth of stock prices, as if the curve would climb forever. Sadly, it never does.
Every boom eventually levels off, and when it does, the money that was riding on the hype goes with it.
That's often when the real work starts though. The people who end up changing the world usually aren't the ones who bet the most during the boom. They're the ones who show up afterwards, pick up what’s left, and figure out how to use those remnants to put the innovation to good, practical use.
The hype builds the infrastructure. Resourcefulness is what makes it last.
What’s different this time?
While the AI boom has striking similarities to some previous cycles in history, there are also some differences this time around.
Unlike the dot-com era, where startups with vague business plans led the charge, the AI Boom is largely driven by large tech giants like Microsoft, Alphabet, Meta and Amazon. These companies fund much of their AI spending with profits from their core businesses (search, cloud, advertising, etc.) which makes them more resilient to a sudden cash crunch than the dot-com startups were. That said, they’re increasingly borrowing to keep up, and that’s one of the warning signs economists are watching.
Digitized Data is also relatively new. By 2007, 94% of the world’s stored information was digital, and it only continues to grow. The data is what modern AI models were trained on, and it’s a big reason this boom took off so fast. They say data is the new oil, and unlike share prices, it doesn’t disappear when the market crashes.
Unlike many dot-com startups, today’s biggest AI companies also have real, paying customers and revenue that’s growing fast.
The money, the data, and the speed of adoption are aspects here that don’t follow a clear historical example. This means that railway mania and the dot-com bubble are useful comparisons, not a roadmap.
So what now?
Knowing all of this, I still don’t have a crystal ball, and I am not an economist or historian. I can speculate, but I can’t reasonably predict how this one will pan out from an economic perspective.
What this does tell me, however, is that when the hype dies down, AI will no longer be just a buzzword being constantly stuffed down our throats. There will be a major shift from a vague “Ooooh AI, it’s revolutionary, let’s put it in everything!” to a more focused “Ok, AI is cool, but how can we really use AI effectively to make our lives better and be worth the investment?”
So in the meantime, I am going to be working on figuring out how to really use AI and make it useful in everyday life and in business. The buildout will bring new work with it, and I’ve started digging into what those new skills and jobs might look like. I’d like to be prepared when the buildout stage hits!
Comments