Back
For models where it made sense, we took some profits on our Asian equity exposure this week. Semiconductor stocks have exploded in value (SK Hynix is up over 800% over last 12 months for example) and part of what we were doing was rebalancing back to where we want to be. But, equally, markets oscillate between greed and fear and it sure looks like greed out there at the moment. Warren Buffett once said “you can’t stand to see your neighbour get rich knowing you’re smarter than he is”. I am sure there are a few people in Korea watching others in two times levered ETFs get rich pretty quickly. The SK Hynix levered ETF is now the largest of its kind in the world (left chart) with foreigners selling to local speculators (right chart):
The last time markets felt this frothy to me was in 2021. One sign back then was a rise in companies selling their stock into strong equity markets. You can see that issuance spike to $115bn in 2021 in the chart below. Well, Google raised over $80bn just this week taking the year-to-date total to close to that already (the chart predates that issuance):
And of course SpaceX and maybe Anthropic and OpenAI are on their way. What to make of all this? As I wrote last week, one comfort you can get from today’s markets is that – even for Korean semiconductor stocks – equity prices have been tracking earnings higher. On a price to forward earnings multiple basis, US equities are actually cheaper than when they started the year:
But, equally, some of today’s AI investor optimism will surely get shaken out in the next few months. But away from the usual ups and downs of equity markets I think it is worth remembering that much of the recent earnings growth you see in the chart above comes from US AI data centre investment. If this starts to slow down or, although it seems hard to imagine today, we build too many then today’s earnings will inevitably turn into tomorrow’s write offs and losses. And, in terms of the rate of change, 2026 is probably the peak of the AI investment boom:
It seems to me that the bet the markets are making is that corporate demand for AI will outpace our ability to build the data centres that power it. If, on the other hand, AI proves harder or more expensive or slower to implement than people think today then the return on some of these data centre investments won’t be that great. Today, my sense is companies are still furiously trying to implement AI and spending budgets are not under pressure. However, technology leaders (including Uber and Microsoft) have recently started to limit their token spending. If this is a start of a broader trend, then some of today’s unbridled optimism will start to unwind. As ever, we shall see, but this is what we will be looking out for at IPS.
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.
