Nick Donato is a financial adviser at Guidance Financial Services in Essendon, Melbourne, helping Australians plan for retirement. You can learn more about our retirement planning services here.
The short answer: possibly, but not in the way the question usually implies. AI could change the economy profoundly while some AI investments still disappoint, because a share price reflects expectations about future profits, not just whether a technology is useful. Nobody can reliably say when a bubble will burst, or which companies will justify their prices. For an individual investor, particularly one approaching retirement, the more useful questions are how much AI exposure a portfolio already has, what is being paid for it, and how a fall would affect their retirement plans.
This article builds on a recent episode of our podcast, where Paul Benson and I worked through ARK Invest's Big Ideas 2026 report. Listen to the full episode here:
One day there's another AI breakthrough. The next, there's a warning that the companies behind it are overvalued.
The first makes people wonder whether they should own more. The second makes them wonder whether they already own too much.
For someone getting close to retirement, that's an uncomfortable place to be. They want their money to keep growing, but they don't want a wave of enthusiasm to put years of saving at risk.
Paul and I recently went through ARK Invest's Big Ideas 2026 report after hearing ARK's founder, Cathie Wood, present it. Some of it struck me as far-fetched. Some of it was fascinating. Paul, as usual, wore the optimist hat, and I wore the pessimist hat.
It gave us both plenty of reasons to be interested in the future. It didn't make either of us want to buy everything associated with it.
Is AI a bubble or not?
It can't be settled by deciding whether AI is useful.
A technology can deliver enormous benefits while investors pay too much for the companies developing it. The internet is the obvious example. It changed almost everything, and many of the companies investors piled into in the late 1990s still didn't survive.
The reason is that a share price reflects expectations about future profits. If those expectations become too ambitious, even a growing business can disappoint. A company can grow its profits strongly, just by less than investors expected, and watch its share price fall while the business itself is doing well.
That's why the choice isn't between believing in AI and believing it's all hype. There's a more sensible position in between.
It's entirely consistent to think AI will be important and still question the price of a particular investment.
Why does ARK think we could be underestimating AI?
ARK's argument goes well beyond better chatbots.
Its Big Ideas 2026 report groups its research into five "innovation platforms": artificial intelligence, public blockchains, robotics, energy storage and multiomics (the analysis of biological data such as DNA, RNA and proteins).
The interesting part is how those technologies might accelerate one another. Better AI makes robots more capable. Cheaper genetic testing produces more health data, and AI can work through that data to speed up medical research.
ARK forecasts that global real GDP growth could reach 7.3% a year by 2030. That's more than double the IMF's forecast of 3.1%, and well above the roughly 3% a year the world has averaged over the past century or so, as ARK's own chart shows.
It's a forecast, not an established outcome, and certainly not a return investors should expect. ARK is an investment manager focused on innovation, and its research is best read with that in mind.
Paul raised a point that doesn't get made often enough. In the US, a large share of recent economic growth has been tied to the AI build-out, yet overall growth hasn't jumped. That suggests AI spending may so far be replacing other activity rather than adding to it. The step change ARK is forecasting might still come. The data doesn't show it yet.
A bullish argument is still worth learning from without adopting every assumption behind it.
Does a growing AI industry mean AI shares will rise?
No. Growth in an industry doesn't automatically translate into good returns for everyone invested in it.
Competition, costs, profits and the price paid all still matter.
A company might spend heavily building the infrastructure everyone needs, then find competition limits what it can charge. Another might build a genuinely useful product and struggle to turn users into lasting profits.
Then there's the question of who ends up with the benefit. A business that uses AI to cut its costs could gain a great deal without ever selling an AI product. The eventual winners may be spread across industries rather than concentrated in the names people associate with AI today.
That cuts both ways. My pessimist's worry on the podcast was that if AI and robots take on large amounts of work in admin, driving, cleaning, factories and the trades, the gains could flow mostly to the people who own the technology. Paul's reply was that history mostly hasn't gone that way. The car put farriers out of work and created far more jobs than it destroyed.
We didn't settle it. But it's a reminder that the economic effects of a technology and the investment returns from it are two different things.
Why is the stock market at record highs?
When we recorded the episode, US and Japanese share markets had just hit record highs, and several European markets were close. That was despite conflict in the Middle East, high fuel prices and rising bond yields. (We've written about what rising bond yields mean separately.)
Part of the explanation is that investors are pricing in a lot of AI optimism. Much of the US market's recent gains have come from a small group of very large technology companies. Markets in Japan and South Korea have been lifted by semiconductor and memory makers that supply the build-out.
The Australian market has largely missed out, because it's dominated by banks and miners rather than technology companies. We've covered that balance in Are You Investing Too Much in Australian Shares?
Record highs aren't proof of a bubble. They do mean a lot of good news is already in the price.
Am I already invested in AI through my ETFs?
Quite possibly. It depends on what the investments hold.
A broad international share ETF or fund can include businesses involved in software, semiconductors, cloud computing and other activities connected with AI. Because global indexes are heavily weighted towards the US, and the US market is heavily weighted towards a handful of large technology companies, the AI exposure in a "diversified" global fund can be larger than people expect.
So someone who has never bought a technology share may already have meaningful AI exposure.
The overlap is where people get caught out. A global ETF, a technology ETF and a few US shares can end up being the same bet held three ways. Owning several funds doesn't necessarily spread risk if they hold the same companies.
The things that tend to reveal it are:
the largest holdings in each fund
how much of the overall portfolio sits in technology businesses
whether a new investment would add to an exposure that's already there
the fees and risks of any specialist fund being considered
Should I change my investments if I'm close to retirement?
An AI headline on its own is a weak basis for changing an investment strategy.
What usually decides it is what the money needs to do. When will work stop? How much will need to be drawn from investments each year? What would happen if markets fell around that time?
Two hypothetical investors can hold the same ETF and be in completely different positions. One hopes to retire in three years. The other won't need to draw on their investments for twenty. A fall means very different things to each of them.
Timing matters more than most people expect. A large fall in the years just before or just after retirement, when withdrawals are starting, can do more lasting damage than the same fall a decade earlier, because assets may have to be sold while prices are down. That's why the mix of investments usually becomes more defensive as retirement approaches. We've written more about this in Will My Money Last in Retirement?
Retirement can also last decades, so avoiding investment risk altogether has its own cost. Money held out of the market can miss years of growth, and moving out after a fall turns a paper loss into a real one. Switching out is also only half a decision. The other half is when to go back in, which means calling the market twice.
We've written more about how much investment risk to take. None of this is something to act on from a podcast or a blog.
How do you decide what to do next?
Before buying more, selling everything or switching strategy, three questions tend to cut through.
What is already owned? Looking through to the underlying holdings, not just the fund names.
What would happen if those investments fell? The effect on spending, on retirement timing, and on the ability to stay invested.
Why is the change being made? A change should have a clear purpose in the overall plan. Fear of missing out deserves attention, but it doesn't establish that an investment is suitable.
I came away from ARK's report interested, though some of it still strikes me as far-fetched. Paul leans towards AI being a genuine paradigm shift, on the scale of railways or electricity. He has never put money into an ARK fund and doesn't plan to, and reading the report didn't change that. Being excited about where a technology could go and being careful about what to pay for it aren't in conflict.
Want a clearer picture of your own exposure?
Not sure how much of your portfolio is riding on one theme? Book an appointment with Guidance Financial Services and we'll work through it together. We don't have our own investment products to promote.
Approaching retirement? Explore our retirement planning services to see how we help people work out what they can afford, how to invest for it, and how to test their plan against different market outcomes.
Common questions
Can AI be a real breakthrough and an investment bubble?
Yes. A technology can be useful and widely adopted while prices for some related investments rise too high relative to the profits those companies eventually earn. That's what happened with the internet in the late 1990s.
Is AI like the dot-com bubble?
There are similarities, including rapid investment, high valuations and a technology that is genuinely changing the economy. There are differences too. Many of today's largest AI-linked companies are already highly profitable, whereas many dot-com companies had little or no revenue. Whether current prices are too high is something nobody can know with certainty in advance.
Is it too late to invest in AI?
Nobody can know in advance whether AI-related share prices will rise or fall from here. Many people who hold a broad global share fund already have exposure to large AI-linked companies. How much additional exposure, if any, is appropriate depends on someone's timeframe, existing holdings and tolerance for volatility.
Do I need an AI ETF to benefit from AI?
Not necessarily. Broad global share funds already hold many of the large companies developing or using AI. Adding a dedicated AI or technology fund on top can increase concentration in the same companies, and concentrated thematic investments tend to be more volatile.
What did ARK's Big Ideas 2026 report predict?
ARK Invest's Big Ideas 2026 report forecasts global real GDP growth of 7.3% a year by 2030, compared with the IMF's forecast of 3.1%, driven by five converging technologies: AI, public blockchains, robotics, energy storage and multiomics. These are one investment manager's forecasts, and they are deliberately more optimistic than consensus.
This article is for educational purposes only and does not take into account your individual circumstances. If you would like tailored advice, we can help you work through the numbers properly.
Forecasts quoted are from ARK Investment Management, Big Ideas 2026, and are not endorsed by Guidance Financial Services.
