Are we in an AI induced asset bubble? Yes, the hallmarks are there but this is still the very early stages and there is so much more to still play out.
Every investment boom has a story.
In the late 1990s, the story was the internet. Businesses were going online, consumers were discovering email and e-commerce, and investors were told that the world had changed forever.
In many ways, they were right.
The internet did change the world. It transformed communication, retail, advertising, banking, media, entertainment, and almost every part of modern business. Many of the companies that dominate markets today were either born during that period or benefited enormously from the internet revolution.
And yet, for many investors, the dot-com boom was a painful experience.
The issue was not that the internet was unimportant. The issue was that investors became too excited, too quickly. They paid high prices for businesses that had little revenue, no profits, weak balance sheets, and in some cases, not much more than a catchy name and a website. When the excitement faded, share prices collapsed. Many companies disappeared altogether.
That history matters today because we are now living through another powerful technology story: artificial intelligence (AI).
AI has captured the attention of investors, businesses, governments, and consumers. It is already being used to write emails, summarise documents, produce code, analyse data, support customer service, generate images, improve search, and help businesses operate more efficiently.
The question for investors is simple but important:
Is today’s AI boom another dot-com bubble, or is it different this time?
The honest answer is that it is both similar and different. History is not repeating exactly, but it may be rhyming.
The dot-com boom was built on real change
It is easy to look back on the dot-com bubble and dismiss it as madness. That would be too simple.
The late 1990s were a period of genuine economic strength. The US economy was growing, inflation was relatively low, unemployment was falling, and productivity was improving. The internet was becoming commercially useful, personal computers were becoming more common, and telecommunications networks were expanding rapidly.
There was a reasonable basis for optimism.
Investors could see that the economy was changing. Businesses were adopting new technology. Consumers were moving online. New companies were being created. Old industries were being disrupted. The future looked exciting.
The problem was that investors took a real trend and stretched it too far.
A good business story became an investment mania. Companies with “.com” in their name attracted huge amounts of capital. Many were listed on the share market before they had proven their business model. Traditional valuation measures were often dismissed as outdated. Profits were treated as optional. Growth was everything.
Eventually, reality returned. Many internet businesses could not generate enough revenue to support their valuations. Capital became harder to access. Investor confidence collapsed. The Nasdaq fell heavily, and many investors who had chased the theme suffered significant losses.
Yet here is the important lesson: the internet still changed the world.
The technology was real. The investment prices were the problem.
Why AI feels similar
There are several clear similarities between the dot-com period and today’s AI boom.
The first is the strength of the narrative. AI is not just another product improvement. It has the potential to change how people work, how businesses operate, and how services are delivered. That makes it easy for investors to imagine very large future profits.
The second similarity is the scale of investment. In the dot-com era, there was enormous spending on fibre-optic networks, telecommunications infrastructure and internet-related businesses. Today, we are seeing huge investment in chips, data centres, cloud infrastructure, energy supply and cooling systems.
Goldman Sachs has estimated that annual AI infrastructure capital expenditure could rise from hundreds of billions of dollars today to well over a trillion dollars a year by the early 2030s. That is a staggering amount of money. It shows how seriously large companies are taking the AI opportunity.
The third similarity is market concentration. A relatively small group of large technology companies has driven a significant portion of recent market returns. These companies are closely linked to the AI theme, either through chips, cloud computing, software, advertising, or data infrastructure.
This does not mean these companies are poor businesses. Many are exceptional businesses. But concentration still matters. When market returns are heavily reliant on a small number of companies, investors become more vulnerable if expectations change.
The fourth similarity is the temptation to bring future benefits into today’s prices. This is where bubbles often form. Investors do not just price what is happening now. They price what they hope will happen over the next five, ten or twenty years. Goldman Sachs just projected that SpaceX’s total revenues will grow from $18.7 billion in 2025 to $474 billion in 2030!
That can be dangerous. The future may arrive, but not as quickly, smoothly or profitably as expected.
Why AI may be different
Despite the similarities, there are also important differences between today’s AI boom and the dot-com bubble.
The biggest difference is the quality of the leading companies.
During the dot-com era, many companies attracting investor money were young, speculative and unprofitable. Some had limited revenue and depended heavily on future capital raisings. Their business models were often untested.
Today, many of the leading AI-related companies are among the most profitable businesses in the world. They have strong balance sheets, large customer bases, established products, global distribution, and deep technical expertise. They are not simply selling a dream. They are already generating large amounts of cash.
That makes today’s environment different from the late 1990s.
The second difference is that AI is already being used widely. Generative AI tools have been adopted quickly by consumers and businesses. People are using them in real time to save time, improve productivity and complete tasks that previously took much longer.
This does not guarantee investment success, but it does suggest that AI is not just a concept sitting in a PowerPoint presentation. It is already inside workplaces, phones, software systems and business processes.
The third difference is that the investment boom is currently more concentrated in established public companies and private infrastructure spending, rather than a wave of small speculative public listings. There may still be pockets of speculation, especially in private markets, but the overall structure is different from the late 1990s IPO boom.
That said, “different” does not mean “risk-free.”
What is a bubble and what are the phases?
An asset bubble is classically defined as a significant and persistent deviation of an asset’s price above its fundamental value driven by speculative forces.
Amundi Investment Institute recently published a report considering the comparisons between the AI boom and the Dot-com bubble (AI Boom or Bubble? Lessons from the Dot-com period, April 2026). Their research highlighted five phases of the Dot-com bubble.
Stage 1 Substitution or displacement stage – Accelerating IPOs, rising valuations and private investments.
Stage 2 Takeoff or boom stage – Market prices increase sharply as investor attention intensifies
Stage 3 Exuberance or Euphoria stage – stretched valuations, soaring trading volumes and substantial participation by retail investors.
Stage 4 Profit taking stage – Some buyers begin to sell out.
Stage 5 Crash – The bubble bursts
Amundi’s conclusion is that the AI boom is currently in stage 2. We are yet to see the euphoria associated stage 3 though we do note the number of large IPO’s coming to the market – SpaceX, Anthropic and OpenAI.
The key question: productivity or overinvestment?
The central issue for investors is whether AI will generate enough economic value to justify the amount of money being spent.
AI could be highly productive. It may help businesses reduce costs, improve decision-making, automate routine work, support employees, and create entirely new products and services. Some research suggests the economic potential could be very large.
However, there is a difference between potential and realised profit.
Businesses still need to work out how to use AI well. Staff need training. Systems need integration. Data needs to be organised. Regulation and privacy issues need to be managed. Some AI tools may save time, while others may create extra complexity. Anyone who has spent twenty minutes trying to get a piece of technology to save them five minutes will understand the problem.
There is also a risk of overinvestment. If too many companies build too much infrastructure too quickly, returns could disappoint. Data centres, chips and power infrastructure are expensive. If demand does not grow fast enough, or if pricing comes under pressure, some of today’s investment may not earn attractive returns.
This is where the dot-com comparison is useful.
In the late 1990s, the world did need internet infrastructure. But it did not need every company, every network and every business model that investors funded. Some infrastructure became valuable over time. Some businesses became giants. Many others failed.
The same may happen with AI. The technology may be transformative, while some investors still lose money backing the wrong companies or paying the wrong price.
What this means for investors
The main lesson from the dot-com bubble is not to avoid innovation. That would have been a costly mistake.
An investor who avoided all technology after the dot-com crash would have missed some of the greatest businesses of the next two decades. The internet created enormous long-term value. But that value did not flow evenly to every company, and it certainly did not protect investors who overpaid at the peak of excitement.
The better lesson is to separate three things:
The technology. The business. The investment price.
A technology can be powerful. A business can be high quality. But the investment can still be poor if the price already assumes everything goes right.
For long-term investors, this means staying disciplined.
It means understanding how much of your portfolio is exposed to a small number of companies or one dominant theme. It means avoiding the pressure to chase recent winners simply because they have gone up. It means recognising that even great companies can suffer sharp falls when expectations become too high.
It also means staying open-minded. AI may well create significant long-term productivity gains. It may reshape industries in ways we cannot yet fully see. Some companies will benefit enormously. Others may be disrupted.
The goal is not to predict every winner. The goal is to build a portfolio that can benefit from innovation without being dangerously dependent on one story.
History does not repeat, but it does teach
There is a famous saying that history does not repeat, but it often rhymes.
The AI boom is not the dot-com bubble all over again. Today’s leading companies are stronger, more profitable and better established than many of the speculative businesses of the late 1990s. AI is also already being used in the real economy, not just imagined.
But the similarities are important. Once again, investors are being asked to price a powerful technology before its full economic impact is known. Once again, large amounts of capital are being committed. Once again, market returns are concentrated in a relatively small group of companies. And once again, there is a risk that excitement about the future leads to overconfidence today.
For investors, the right response is not panic. It is discipline.
AI may be one of the most important technologies of our lifetime. But as the dot-com period reminds us, being right about the technology is not the same as being right about every investment attached to it.
The internet changed the world, but not every internet stock was a good investment.
AI may do the same.
The sensible path is to remain invested, stay diversified, focus on quality, avoid unnecessary concentration, and remember that valuation still matters.
In other words, enjoy the innovation, but do not let the excitement drive the portfolio.
The future may be exciting, but good investing still requires patience, balance and a clear head.
We may be at the start of a new earnings bubble, but we are yet to see that flow through to broad-based exorbitant pricing. While we don’t know what will burst the bubble, I suspect that there is still a lot of air to go into this bubble.
Andrew Aylward is Chief Investment Officer at Keep Wealth Partners.
For more information contact us on 03 8610 6396
Keep Wealth Partners Pty Ltd (AFSL 494858)
This information is of a general nature only and may not be relevant to your particular circumstances. The circumstances of each investor are different, and you should seek advice from a financial planner who can consider if the strategies and products are right for you.
