One of the first rules of putting together a portfolio is that you don’t put a whole lot of your apples into the one basket. Concentration creates risk.
So, it can be unsettling to look at just how big technology companies have become in global share markets. The IT sector is close to one-third of the global index, and if you include companies like Amazon, Google and Meta, which are categorised into other sectors, it’s more like 40 per cent.
To make things even less comfortable, if you’re worried about tech, or AI, or semiconductors being in a bubble, as seems to get discussed on a daily basis, it can leave you wondering what to do with your own portfolio’s exposure.
Some context
Before you get too concerned about the grim warnings of share markets having unprecedented concentration in technology companies, just stop and think about the role technology plays in the world today compared to 30 or 40 years ago. It is all pervasive.
From the stuff that’s obvious in our day-to-day lives, like phones that rarely leave our side or the complete reliance we have on the internet for information, communication and even shopping, to the stuff that’s not so obvious, like the reliance our critical businesses and institutions have on technology, it should be expected that would result in a huge increase in tech’s market presence.
The AI tsunami
It’s estimated that global spending on AI capex will be more than US$1 trillion in 2026 and will increase to about US$1.2 trillion next year.
In the US alone, spending this year will be around US$750 billion. That’s a massive injection of money that doesn’t just disappear into a black hole, it bounces around the economy paying for computer hardware, electrical equipment, building materials for the AI data centres, contractors to do everything from earthworks to installing the lighting and ventilation gear.
Then those companies spend money on their own inventory and wages, and so it goes on.
To 27 July, the S&P 500 is up 8 per cent for the year to date, but boring old industrials have risen 18 per cent, by comparison, the IT sector is up 15 per cent. Within that sector, construction machinery and heavy trucks, both vital to building data centres, have gone up by 45 per cent, and electrical components and equipment by 20 per cent.
Since the start of 2024, monthly durable goods orders (excluding aircraft), have risen from US$74 billion to US$85 billion, with new orders rising at an annualized pace of almost 13 per cent.
It’s no wonder US companies are on target to report year on year earnings growth of about 25% for the second quarter.
But will it last?
The obvious question on every smart investor’s lips is: those numbers are really impressive, but will they last?
The so-called hyperscalers, those companies responsible for the headlong rush to build AI capacity, like Amazon, Google and Microsoft, have all but burned through their free cash. In 2024 they raised about US$15 billion in debt and equity, in 2025 that jumped to more than US$100 billion, and so far this year, it’s more than US$300 billion!
That’s a striking increase, but you’re talking about some of the biggest companies in the world, each of which has barely begun to stretch their balance sheets.
What about all the talk of bubbles?
There’s no universally accepted definition of a bubble, so it’s not like you can just compare today’s market against some measurement and declare it to be one.
When you talk about a tech bubble, most seasoned investors immediately think of the dotcom boom at the end of the twentieth century. In early 2000, at its peak, the average forward price to earnings (PE) ratio of the top four tech companies was 89x. Today, it’s 24x.
The forward PE for the tech sector today is 22x, in 2000 it was 55, 2.5x more expensive!
If you want to sound really smart, you say, “Yeah, but it’s not a valuation bubble, it’s an earnings bubble.”
But, as usual, the market is way ahead of you. The pinup child for the AI bubble theme is the Korean stock market, which has risen by about 152 per cent since the start of 2025. It’s on a forward PE of 6.2x.
Samsung Electronics, which has risen by more than 300 per cent since the start of last year, is on a forward PE of 4.8x.
Micron, the crown jewel of the semiconductor shortage, saw its earnings increase by more than 1,200 per cent since the start of last year, yet over that time the share price has risen about 850 per cent. Analysts expect earnings to grow nearly 60 per cent over the next year, but as of July 21 it was trading at just 6x earnings.
Compare that to the Australian market, which is trading on about 18x next year’s earnings, which are projected to grow by about 8 per cent.
In other words, the market is telling you it doesn’t believe those tech earnings will keep growing the way they have been.
Some more context
On top of global demand for data centre capacity potentially tripling by 2030, Bank of America forecasts annual shipments of one million humanoid robots by 2030, with each one containing upwards of US$13,000 worth of computer hardware. One market forecast is that the robotic semiconductor market will grow from US$11.2 billion in 2025 to US$41.2 billion by 2030.
Also, the rise of frontier AI is likely to accelerate enterprise hardware refresh cycles over the next five years, driven by the need for AI-capable compute and hardware-based security features.
The great broadening
To 28 July, Microsoft has fallen 29 per cent from its recent high, Google’s down 18 per cent, Meta by 25 per cent, Tesla by 38 per cent and Amazon and Nvidia are both down 17 per cent.
When the biggest companies in the market are in drawdowns like that, most people would presume the market will have taken a bath. But the S&P 500 is only 2.5 per cent from its highs.
The reason it hasn’t suffered a dramatic fall is because so many of the other companies are seeing their share prices go up, which is reflected in the equally weighted S&P 500 (that is, every stock has a 0.2% weighting), which is at all-time highs.
It’s what’s referred to as a broadening of the market, and it’s a very positive thing.
What do you do?
Markets never, ever go up in a straight line. And when you hear about Korean and Taiwanese retail investors making overnight fortunes by taking out margin loans and buying triple levered TSMC ETFs, you know there’s going to be tears.
The Korean index, the KOSPI, has already fallen an eye watering 34 per cent from its recent peak.
But the AI tsunami is spreading across global markets, making it difficult to find a sector that’s untouched.
It appears the market is already pricing in a slowing of earnings growth in the red-hot areas of tech and is turning its attention to the parts of the market that are either benefiting from the AI spending or companies that are finding ways to use AI to increase their productivity and earnings growth.
Something you should never do in times like this is swing for the fences. As tempting as it might be to load up on the hot stocks, remember the apples and baskets and don’t get greedy. The diversification offered by index funds and ETFs is one way to mitigate risk.
You have to be able to sleep at night, and nobody’s gone broke taking profits, but it’s worth recalling that Alan Greenspan, then the head of the US Federal Reserve, said in December 1996 that the share market was showing “irrational exuberance”. The S&P 500 more than doubled over the following three years.