The Viewpoint: AI: It is and isn’t a bubble

The most frequently asked question I get in client meetings: Is AI a bubble? Markets are becoming an AI singularity. It is the main, and some investors say only, source of both investment performance and economic growth.

While it is an exaggeration to say markets and economies are becoming an ‘all-in’ bet on AI, it is perhaps heading in that direction. Most indices, both developed and emerging market, are dominated by AI winners, meaning passive investors are not as diversified as they used to be. Equally active managers are being required to take on more AI exposure to meet, let alone beat, their benchmarks. It’s because of this the bubble question is asked.

I have seen two equity bubbles in my investment career. The first was the internet bubble of 1999. The second in US housing in 2007. They are rare things. Bubbles are talked about much more than they occur. Indeed, the only difference between a bubble and a bull market, as someone astutely said, is if you are invested in it! What these two bubbles have in common is this idea of a singularity: by their end, both the stock market and the economy were exceptionally vulnerable to their key driver disappointing. As the US housing market collapsed, and the internet boom turned to bust, it catalysed significant bear markets.

The internet bubble

My memories of the internet bubble, which is the one most relevant to today, are numerous. The first memory is the time between when I thought it had become out of hand – and reasonable investors would rightly exit – and its end was short in months but high in share price moves. Between October 1999 and December 1999, an eight-week period, the Nasdaq index rose nearly 50%. This was after it had already risen 90% in the 12 months before! The same index has risen nearly 30% since the end of March 2026 having risen 17% in the 12 months before. At least on this precedent we are not in a bubble.

What I also remember is that even at the peak of the internet bubble, the most euphoric investors underestimated the impact it would have on society and its usage levels. What these same investors also got wrong however was the economics of it. Revenue growth came first; profits did not always follow later. Another way of saying this is that selling loss-making products is easy, selling profitable ones is much more difficult. Revenue growth always looks exponential in the early stages. The sustainability of it and associated profits take longer to understand. The ultimate economics of AI usage may ultimately be strong. But for now, they are largely unknown.

Selling loss making products is easy, selling profitable ones is much more difficult.

Another memory is how wrong assumptions of who were the winners and losers in the early stages of the internet proved to be. AOL (America Online) was the king that wasn’t. At its peak, it was the undisputed gatekeeper of the internet for millions of Americans. It used its inflated stock to acquire Time Warner, leading to a $99bn write-down in 2002. Webvan was an online grocery delivery startup that went public in 1999. It is a great example of where the excitement over a new product was correct, but the economics of it resulted in it collapsing into bankruptcy in 2001. Businesses such as Ocado are still struggling with the economics of this industry today.

Equally businesses which were written off did well. The most obvious example of this is Apple. In 1997, Apple was viewed as an internet loser. It had no internet strategy, declining market share, and was weeks away from bankruptcy. Michael Dell famously remarked that if he ran Apple, he would "shut it down and give the money back to the shareholders." Investors viewed Apple through the lens of the 1990s desktop PC wars, which Microsoft had won. They failed to realise that under a returning Steve Jobs, Apple would build the hardware ecosystem (the iMac, iPod, and ultimately the iPhone) that turned the internet from a static desktop experience into a mobile utility.

The internet created huge winners. The winners were not however always the companies investors initially thought they would be. The mantra of first mover advantage was fundamentally wrong: right-mover advantage mattered more. Most relevant for the AI debate is the view that pure internet plays (similar to AI native companies today) would kill legacy bricks-and-mortar (incumbents). In the end omnichannel (internet and bricks-and-mortar) won with even Amazon becoming an omnichannel company. AI native companies are winning today, but it may be the incumbents who use AI who win in the end.

In summary, while AI is clearly driving a large share of equity market and economic optimism, current market moves do not yet match the late-stage excesses of 1999 and the internet boom. More importantly, the lesson from the internet era is that investors often correctly identify a transformative technology but misjudge its economics and eventual winners: early leaders can fade, while overlooked incumbents or later movers may prove more successful. In that sense, AI may well reshape the economy profoundly, but uncertainty remains around which business models will generate durable profits and whether today’s apparent winners will still lead in the long run.

The lesson from the internet era is that investors often correctly identify a transformative technology but misjudge its economics and eventual winners

An open mind is key

Where does this leave investors? If the best risk tool is a pair of eyes, the best reward tool must be an open mind. There is an unusually wide range of outcomes for markets and companies within them over the coming years. It is highly unlikely that any portfolio will not have to adjust to events as they occur. Interestingly, one of the leading venture capitalists in the AI space, US-based a16z, is enacting a strategy of broad diversification of their investments as they too do not have certainty over who will win. They are also more optimistic about the role of incumbents than listed equity markets are. It would appear they have learnt the lessons of the internet era, and chosen breadth over certainty, and a balanced view of AI native companies versus incumbents. This is a good approach for all investors.

From my perspective, witnessing what may or may not be the third bubble of my career forming, my answer to the question ‘is this a bubble?’ is that there is no value in answering yes to that. Mentally, once that position is taken, the ability to have an open mind reduces significantly. No one has privileged access to the future, and investing is probabilities not certainties. There will always be a probability that AI is and isn’t a bubble. This means productive time is spent considering both scenarios and building portfolios around them. As the future reveals itself, whatever that may bring, it will mean being better placed to adapt to it.

There will always be a probability that AI is and isn’t a bubble.

For professional investors only.  This material is not suitable for a retail audience. Capital at risk. This is a financial promotion and is not investment advice. Past performance is not a guide to future performance. Reference to any security is for information purposes only and should not be considered a recommendation to buy or sell.

The value of investments and any income from them may go down as well as up and is not guaranteed. Investors may not get back the amount invested. Portfolio characteristics and holdings are subject to change without notice. The views expressed are those of the author at the date of publication unless otherwise indicated, which are subject to change. Forward looking statements are subject to certain risks and uncertainties. Actual outcomes may be materially different from those expressed or implied.