How exposed are asset owner portfolios to AI?
The question of whether an AI bubble is about to pop has been debated for some time, but which parts of their portfolio should asset owners be focusing on? Research suggests more than they might think (Seong Joon Cho/Bloomberg)
Good morning. Is the artificial intelligence bubble about to pop? This is a question we’ve been asking for a long time.
I certainly remember this question being discussed in the previous publication I edited in the ancient history of early 2024.
There have certainly been jitters lately. South Korea’s Kospi index, which was previously having a great time, has dropped 30 per cent since June in part due to troubles at chipmakers SK Hynix and Samsung. Meanwhile valuations, measured by the Shiller price-to-earnings ratio, are higher than they’ve been since 2001. Plus the recent, erm, troubles faced by the Situational Awareness hedge fund haven’t helped the mood.
But on the plus side the S&P 500 has stopped its vertiginous climb and has been quietly moving sideways for months.
In the meantime, what should asset owners be focusing on?
One answer to this question is understanding how much exposure to AI they actually have, according to Frank Benham, Meketa Investment Group’s director of research.
Much of the discussion around AI exposure focuses on public equities such as Nvidia - and maybe private credit if you’re worried about data centres.
But this graph, which Benham put together, highlights why AI is a problem across a multi-asset portfolio:
Benham said: “I suspect the exposure in infrastructure is the exposure most people may miss, because it is a second-order linkage. That strikes me as an area that may not be as obvious to people.
“If asset owners are only looking at their public equity exposure then they may have more exposure to AI than they think. Even if they are only looking at their public and private equity exposure.”
Is this proliferation of AI exposure driven by demand or supply (i.e. are asset owners seeking it out, or are asset managers giving it to them without being asked)? Benham said it probably depends on the asset class.
He said: “Private equity, particularly [venture capital], that’s been all about AI and that’s where the money has been going. There’s clearly that side of people seeking AI exposure deliberately.
“On the flip side I would hypothesise that many people investing in private credit are not
looking for AI exposure. They are looking for high yields with stable income and returns.”
Understanding where a portfolio sits in terms of its AI exposure is a question asset owners need to answer before asking what they should do about a potential AI bubble, Benham said.
He said: “A good starting point is to take a step back and look at your portfolio allocation.
“Maybe instead of modifying your equity portfolio you have other assets which can provide you
with ballast. Something which may help you with protection in a downturn like Treasury bonds.
“I think over the next few months we will have a lot of conversations with our clients about AI exposure. If OpenAI and Anthropic were to go public later this year that could theoretically increase the concentration even more.”
