Back
I have been doing this job for 16 years now. Early on, when the US started to out-perform, I confess to thinking that the US was expensive (in the early 2010s!) and there would be some sort of reversion to the mean. Well I, and all the other value investors and technology bears and grumpy Europeans, were wrong. The US was in fact cheap in the early 2010s (even with much higher prices and p/e ratios) because US technology companies from that era became some of the greatest money-making machines ever created. Meta (then Facebook) had its IPO in 2012 with $3.7bn of revenue and $0.7bn of profits. Today those numbers are $165bn and $62bn. Microsoft and Google combined generate almost $200bn in profit each year.
I mention this because this because we had a client question this week on whether we are in an AI bubble (and what we would do about it if we were). And, for me, this earnings story is the key difference between today and the dotcom bubble we saw in the early 2000s. Investors back then were of course completely right that the internet would change everything. It’s just the product (and revenues) weren’t there yet. It was a jam tomorrow story and the jam took a lot longer to arrive than the 1990s techno-optimists expected. Equity prices are high today (and on for another solid quarter in Q3) but this time the earnings story is there for the high-flyers. Earnings for the US Magnificent Seven technology giants grew at 26% year-on-year in the second quarter, compared to 4% for the rest of the US market.
I have written plenty of times here that US technology giants are a juggernaut that I do not want to stand in the way of. The trend in the chart above is in place until proven otherwise. But large companies cannot grow at faster than nominal GDP forever. At some point this too will pass. What are the risks when it does?
The real economic problem from the 2000s dotcom bust came not from the stock market but from the $500bn in fiber-optic cable investments made at the time. These have ultimately proved to be incredibly valuable (think of all that Netflix and working from home during Covid for a start). But in the early 2000s there was massive over-capacity and much of the debt used to finance these cable investments was not paid back. This was a classic over-investment boom and bust and it bled into the real economy triggering a (mild) recession. US equity markets halved in 18 months and those value investors and technology bears I mentioned at the start had their moment in the sun. (Being very right in the early 2000s probably meant you were very wrong in the early 2010s. Such are the challenges of this job).
There are obvious parallels here to the $461bn the Magnificent Seven expect to invest next year in (mainly) data centres. Also, this week Nvidia announced a $100bn investment in OpenAI. This looks to me like a version of giving your customers money to buy your product. Vendor financing like this was part of the fuel for Cisco’s dramatic boom and bust in the dotcom era. The over-investment risk looks real to me. But how might it play out?
One way is that (as in the 2000s) demand would come through much more slowly than expected leading to chronic data centre over-supply. This is of course possible but the demand side looks pretty solid to me. Already 10% of the world’s population use OpenAI on at least a weekly basis and plenty of these people are prepared to pay for it. Instead, I think the risk is that people (and models) continue to get smarter. China’s Deepseek model (which was not developed on Nvidia’s latest chips) has showed there may be smarter and cheaper ways to run AI models. Clever people often find ways round resource constraints. As technology moves on, it may be that today’s data centres have a much shorter shelf life than those making the investments today expect.
If we do in fact get another over-investment technology boom and bust, here is what happened to US base rates in the 2000s:
I am not sure anyone is smart enough to know when and if this AI technology boom will end. But if (like me) you think over-investment is a real risk here then bonds (which needless to say benefit from falling interest rates) look like a good hedge. And the fact that today’s yields are comfortably above inflation mean your hedging is also making you money. This is one good reason we remain overweight bonds, where appropriate, in our portfolios today.
Chris Brown, CIO
cbrown@ipscap.com
The value of investments may fall as well as rise and you may not get back all capital invested. Past Performance is not a guide to future performance and should not be relied upon. Nothing in this market commentary should be read as or constitutes investment advice.
