If you’re a history buff like me you’ll know that, through the centuries, markets have been prone to cycles of euphoria and despair, boom and bust. 

The Dutch tulip mania of the 1600s, the South Sea Bubble of the 1700s, Japan’s asset bubble in the 1980s and, more recently, the dotcom crash of the late 1990s are all examples. 
Each of these episodes carried familiar hallmarks: rapid capital inflows, inflated valuations and a widespread belief that this time just might be different. 

A typical bubble begins with ‘smart money’ investors spotting an opportunity early. As optimism grows, institutional investors pile in, followed by the broader public, often late to the party. Prices rise well beyond fundamentals, creating a self-reinforcing cycle of greed and momentum. Eventually, reality catches up, expectations fall short, and the bubble bursts.

Today, artificial intelligence (AI) has become the defining theme of global markets. Compared to roughly 48,000 publicly listed companies worth $124 trillion globally, there are just 775 firms that are primarily AI-focused which account for a staggering $19 trillion in market value.¹

This level of concentration in such a narrow set of companies is striking.

Comparisons with the dotcom era are natural. Between 1995 and 2000, the tech-heavy NASDAQ index soared on the promise of the internet, only to collapse when revenue and profit growth failed to match expectations.²   

In the current cycle, AI has taken on a similar aura - a transformative technology said to revolutionize industries, drive productivity, and underpin future economic growth.

There is also massive capital expenditure flowing into AI infrastructure. Historical US capex data shows that, as a proportion of US GDP, AI spending today exceeds that of the telecommunications / fibre boom in the early 2000s – although it’s still currently less than half of what was spent on expanding railroads in the mid-1800s.³ 

That tells me investors and companies alike are betting big on AI’s promise, and will continue to do so.

My answer is: yes, but with some important caveats.

The valuations we see today clearly suggest bubble-like characteristics. Some AI firms are priced as if they will dominate the future economy, despite many having unproven or early-stage business models. This disconnect between hype and reality is a classic warning sign.

But unlike the dotcom boom of the late 1990s, when a flood of companies went public and drove markets higher, today’s AI startups can stay private for longer given the constant investment from venture capital funds, sovereign wealth funds, family offices and other tech investors.

And bubbles are not entirely bad. They emerge because there is often truth at the core of the story. The dotcom bubble burst but, two decades later, the internet dominates how we learn, work and play. Similarly, the railroad boom of the 19th century left investors nursing heavy losses, but it also laid down infrastructure that reshaped economies.

AI may follow a similar trajectory. Even if valuations correct sharply at some point between now and, say, 2030, the underlying technology is likely to remain transformative. Productivity gains from AI could be real and lasting, even if investors misprice the timeline and scale of adoption in the short term.

The critical question is not whether AI is a bubble – the signs are most certainly there - but how and when it ends. Market cycles don’t end neatly. They tend to unravel when expectations exceed what can realistically be delivered. In AI’s case, that could mean years of disappointment if applications take longer to commercialise, or if regulation and costs temper the rollout.
At the same time, there’s a lot of money out there waiting to be invested: since Covid, US household cash levels have surged by over 30%.⁴ With global liquidity abundant and central banks aiming to stimulate growth in many regions, speculative fervour can last longer than most expect.  Timing the top of a bubble is notoriously difficult.

AI may well be the defining technology of our age, but investors should remember that markets often overshoot. Yes, AI is in a bubble. Yes, prices will probably correct. But when the dust eventually settles, AI’s long-term contribution to productivity and growth will remain.

For investors, the lesson is clear: participate, but with caution. Avoid chasing hype and focus instead on fundamentals, diversification and risk management. The bubble may well burst, but AI - much like the internet before it - is here to stay.

1 visualcapitalist.com and stockviz.com
2 Bloomberg
3 Microsoft, Forbes, Bloomberg, Reuters
4 BoFA Research Investment Committee: Federal Reserve Flow of Funds

Information is current as at 4 September 2025. Pie Funds Management Limited (“Pie Funds”) is the issuer and manager of the funds in the Pie Funds Management Scheme and the Pie KiwiSaver Scheme (“Schemes”), the product disclosure statements of which can be found at www.piefunds.co.nz. Any advice is given by Pie Funds and is general only. Our advice relates only to the specific financial products mentioned and does not account for personal circumstances or financial goals. Please see a financial adviser for tailored advice. You may have to pay product or other fees, like brokerage, if you act on any advice. As manager of the Schemes, we receive fees determined by your balance and we benefit financially if you invest in our products. We manage this conflict of interest via an internal compliance framework designed to help us meet our duties to you. For information about how we can help you, our duties and complaint process and how disputes can be resolved, or to see our disclosure statement, please visit www.piefunds.co.nz.  Please let us know if you would like a hard copy of this disclosure information.