Diversification in the Age of AI

When different investments may share more of the same risk than you think

If you follow the financial news, you've probably encountered some version of the same question lately:

Are we in an AI bubble?

Earlier this year, in our 1Q26 Newsletter, we took a closer look at the extraordinary investment boom surrounding artificial intelligence and the questions it raises for investors. At the time, our conclusion was intentionally measured. Artificial intelligence is very real, its potential economic impact is enormous, and many of the companies leading its development are highly profitable businesses. At the same time, history tells us that transformative technologies can attract more investment, and higher expectations, than even extraordinary future growth can ultimately justify.

Those questions haven't gone away. If anything, the enormous amount of capital flowing into AI has made them even more relevant.

We'll return to that broader discussion in our fourth-quarter newsletter, when we'll take a deeper look at market valuations, AI capital spending, credit markets and perhaps most importantly, the massive upcoming IPOs of AI's two biggest players: OpenAI and Anthropic.

For now, though, we think there is another important lesson emerging from the AI boom:

Diversification isn't just about how many different investments you own. It's about how many different risks you own.

When Different Investments Share the Same Story

Most investors understand the basic idea behind diversification. Rather than placing all of your money in one company or one type of investment, you spread it among stocks, bonds, real estate, infrastructure and other investments. When one part of the portfolio struggles, another may behave differently and help cushion the impact.

That principle hasn't changed.

But today's AI investment boom offers a great example of why simply owning different types of investments doesn't necessarily mean they are being driven by different economic forces.

The connection is easiest to see in the stock market. Major U.S. indexes have become increasingly influenced by a relatively small group of very large technology companies, many of which sit directly at the center of the AI boom.

But the story doesn't stop with stocks anymore.

Building the infrastructure required for artificial intelligence requires an extraordinary amount of capital. Technology companies are spending hundreds of billions of dollars on semiconductors, servers, data centers, networking equipment and the electrical infrastructure needed to power it all.

And that money has to come from somewhere.

Increasingly, the AI buildout is being financed not only through corporate cash flows and public stock markets, but through corporate bonds, private credit, real estate financing, infrastructure investments and other parts of the capital markets.

That creates an interesting diversification question.

Imagine a portfolio containing:

• A broad U.S. stock index

• Investment-grade corporate bonds

• Private credit

• Infrastructure investments

• Real estate or data-center investments

• Utility companies

On paper, that may look like exposure to several very different areas of the market.

Look underneath the labels, however, and you may find that several of those investments share exposure to the same underlying assumption: continued massive investment in artificial intelligence infrastructure.

The stock portfolio may own the technology companies building AI systems. The bond portfolio may lend money to those same companies. The infrastructure portfolio may finance data centers. The utility investment may depend partly on rapidly growing electricity demand from those data centers. Private lenders may finance still other pieces of the same buildout.

These are not identical investments, and they won't necessarily behave the same way. A highly rated corporate bond is very different from a technology stock, for example.

But they can still share exposure to the same economic driver.

And that distinction matters.

The Internet Really Did Change the World

History provides a useful analogy.

The technology enthusiasm of the late 1990s is often remembered simply as the "dot-com bubble." But one of the most important lessons from that period is easily overlooked:

The investors were right about the technology.

The internet really did transform commerce, communications, entertainment and nearly every other part of modern life. In many ways, its eventual impact exceeded what even the optimists of the late 1990s imagined.

That did not mean every internet-related investment made in 1999 was a good investment.

A transformative technology can be enormously successful while investors simultaneously overpay for some of the companies, infrastructure or financing associated with it.

The same could ultimately prove true of artificial intelligence.

AI may become one of the most important technological developments of our lifetime. That doesn't automatically tell us whether every company, data center, bond, private loan or infrastructure project being financed today will produce an attractive investment return.

Those are two very different questions.

So, Is AI a Bubble?

Maybe parts of it are. Maybe not.

We don't think investors need to make that binary prediction in order to build sensible portfolios.

The better question is this:

If expectations surrounding AI changed dramatically tomorrow, how much of your portfolio would actually be affected?

That is where diversification becomes more interesting.

True diversification isn't simply a matter of owning ten funds instead of five, or combining stocks with bonds and alternatives. It requires understanding the economic risks underneath those investments and making sure that seemingly different pieces of a portfolio aren't all dependent on the same outcome.

None of this means we believe investors should avoid artificial intelligence.

Quite the opposite. Innovation has always been one of the great engines of economic growth, and investors should participate in it. AI may create extraordinary opportunities for businesses and investors for many years to come.

The objective isn't to eliminate AI exposure.

The objective is to understand it, make it intentional, and size it appropriately.

What We're Watching

At Firelands Wealth Management, we're increasingly looking beyond the traditional labels attached to investments and asking what is actually driving them.

Where does our technology exposure come from? How concentrated are broad stock indexes? How much AI-related borrowing is entering the corporate bond market? How much of the growth in private credit, infrastructure and data-center investment ultimately depends on the same capital-spending cycle?

Those questions don't mean diversification has stopped working.

They mean good diversification requires looking beneath the surface.

We'll take a much deeper look at these issues in our fourth-quarter market commentary, including the extraordinary level of AI capital spending, current valuations, developments in the credit and bond markets, and the broader economic backdrop as we look toward 2027.

For now, the takeaway is simpler:

Diversification still works. But knowing what you own is only the beginning. Understanding what is actually driving those investments may matter even more.