You may have heard (or read) that there is speculation about whether there is an ‘AI bubble’ in the investment markets at the moment. Some people are comparing the current situation with the ‘dot.com bubble’ at the turn of the century, but we believe that this time it could actually be different (but there are still risks in areas of the markets).
This article is designed to give you a framework to think about whether there is a bubble – and help you with understanding the varying messages you may hear in the headlines.
Background
Over the last three years (since the launch of ChatGPT) and more particularly in the last 12 months, Artificial intelligence has moved rapidly from a technical topic into a dominant investment theme.
Company announcements, market commentary, and share prices increasingly reference AI as a driver of future growth. At the same time, a small group of large technology companies now represents an unusually large share of global equity markets.
Given the rapid growth, it is reasonable to pause and ask whether this enthusiasm reflects long-term opportunity, short-term excess, or a combination of both.
From a Moneyworks perspective, the question is less about predicting market turning points (which would be looking at timing the markets, which generally isn’t feasible) and more about understanding where risks may be building and how they could affect diversified, long-term portfolios.
Technology change and investment risk are not the same thing
AI is already being used in practical ways across many industries (the February blog article will outline Carey’s personal experience in ‘mastering AI’).
Professionals are using it to support research, analyse information, draft documents, assist with medical imaging, improve logistics, and streamline administrative tasks. For many businesses, RPA (Robotic Process Automation) – which is what Moneyworks have built with Millie over the last 7 years, is often grouped under their AI suite of solutions.
These changes appear real and are likely to continue.
Investment markets, however, do not price today’s usefulness.
They price expectations about future earnings. When a powerful new technology emerges, markets often assume rapid adoption, strong pricing power, and high profitability across a wide range of companies.
History suggests that these assumptions tend to be overly optimistic, particularly in the early stages.
Recognising the difference between technological impact and investment outcomes is critical for investors who care about long-term results rather than short-term excitement.
Where the financial risks may be building
If there is bubble-like risk in the current AI cycle, it is less likely to sit with the technology itself and more likely to sit with how it is being financed.
AI requires extraordinary levels of capital.
Data centres, advanced semiconductors, energy supply, cooling systems, and ongoing upgrades all demand significant investment. This has led to an infrastructure build-out that, in many cases, is being funded with increasing levels of debt.
The largest technology companies are better positioned to manage this (eg Alphabet (Google), Meta (Facebook), Amazon, Microsoft.)
They generate substantial cash flow and can fund much of their investment internally. For them, the risks relate more to execution and competition than to financial survival.
Smaller companies and infrastructure providers face a different challenge. Borrowing heavily to fund long-lived assets in a rapidly evolving technological environment introduces uncertainty.
When technology changes quickly, it becomes harder to predict how long expensive infrastructure will remain economically useful.
When businesses borrow debt to finance their investments, it is assumed that the environment is stable. Fast-moving technology rarely offers that.
Rapid change creates uneven outcomes
AI capabilities are evolving at an extraordinary pace. New models and tools emerge every few months, often rendering previous approaches less competitive. (For example, Marloo – our financial services specialist note taker and transcriber is issuing improvements and developments almost weekly as part of it’s buildout).
From an innovation perspective, this is positive. From a business and investment perspective, it introduces uncertainty.
Many organisations are still experimenting with AI rather than fully integrating it into how they operate. (An excellent podcast to learn what companies are and are not doing can be found at Business Desk – The Business of Tech https://businessdesk.co.nz/business-of-tech).
Productivity gains are uneven and often difficult to translate into sustainable profit growth. In competitive markets, efficiency gains may benefit customers through lower prices rather than shareholders through higher margins.
Markets have a long history of extrapolating early success too far into the future. This does not mean AI will fail to deliver value, but it does suggest that returns are unlikely to be smooth or evenly distributed.
Employment change has investment implications
AI is also reshaping the nature of work. Unlike earlier waves of automation, which largely affected manual or repetitive roles, AI is influencing professional and knowledge-based work. This has implications for labour markets, training, and income distribution.
Over time, economies tend to adjust. In the shorter term, periods of transition can be disruptive.
Political responses, regulation, and public expectations all influence how gains and costs are shared. The potential changes have raised the issue more than once of the future necessity of a UBI (Universal Basic Income).
For investors, these second-order effects matter. They can affect company costs, regulatory settings, and long-term profitability, even when the underlying technology is sound.
Lessons from earlier technological cycles
History offers useful perspective. Railways, electricity, telecommunications, and the internet all attracted intense investment early on.
Many companies failed, and many investors overpaid. Yet the technologies themselves transformed economies and improved productivity over decades.
The lesson is not to avoid innovation, but to recognise that being right about long-term change does not automatically lead to good investment outcomes. Valuations, balance sheet strength, and competitive advantages remain critical.
What this means for long-term investors
For most investors, the sensible response is neither enthusiasm nor avoidance, but balance. Many diversified portfolios already have meaningful exposure to AI-related companies through global equity markets. Concentration risk, particularly in large technology firms, deserves attention.
A disciplined investment approach continues to focus on fundamentals. Cash flow matters. Financial resilience matters. Diversification matters. Ethical and responsible investors also need to consider energy use, labour impacts, and governance as part of assessing long-term sustainability.
The greatest risk may be assuming that rapid technological progress guarantees strong investment returns, rather than recognising that markets often price optimism well ahead of reality.
Conclusion
AI is likely to play an important role in shaping the global economy over the coming decades. It will change how businesses operate and how work is done. That does not mean every company associated with AI will succeed, nor that current market valuations will always be justified.
For long-term investors, steady decision-making, diversification, and a focus on underlying fundamentals remain as important as ever, even during periods of technological excitement.
We monitor our fund managers to understand the investment approaches they are taking and the kind of exposures to the ‘AI bubble’ that they may have. We continue to monitor how our fund managers are approaching AI-related opportunities and risks, and how those exposures sit within well-diversified portfolios over time
