85% of family offices prioritize AI investments, 75% foresee a bubble

1 hour ago 18

A new RBC Wealth Management and Campden Wealth survey finds North American family offices betting big on AI while bracing for a crash within five years

North America’s wealthiest families are all in on artificial intelligence. They also think it’s going to blow up in their faces.

A new survey from RBC Wealth Management and Campden Wealth found that 85% of family offices rank AI as a top investment theme. Meanwhile, 75% expect an AI investment bubble to burst within five years.

The numbers behind the AI paradox

The findings come from the 2026 North America Family Office Report, which surveyed 155 family offices. The participating families have an average net worth of $2.25 billion.

Of those surveyed, 85% said AI is a top theme likely to generate returns over both the short term and the longer term. The report defines the short term as the next 12 months and the longer term as two to five years.

An even more pointed figure: 59% picked AI as the leading investment theme for 2026. Not one of several. The one.

Then comes the twist. Three-quarters of respondents warned that an AI investment bubble may burst within the next five years.

Adam Ratner, the report’s research director, has a name for this tension. He calls it “the AI paradox.”

Old-school assets win the long game

The paradox shows up most clearly when the survey stretches the time horizon. Looking two to five years out, family offices actually favor a more traditional option.

On that longer outlook, 69% of respondents prefer US equities. AI comes in second at 64%.

Caution, however, is not the same as retreat. A notable 41% of family offices plan to increase their AI exposure specifically within private-market technology holdings.

Riding high after a strong 2025

Some of this confidence traces back to recent performance. According to the report, 84% of family offices expect their direct private equity investments to match or beat 2025’s results.

Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.

Read Entire Article