AI leads family office investment priorities: JPMorgan

JPMorgan’s survey reveals that 65% of 333 family offices across 30 countries prioritize AI as their top investment theme, though actual allocations remain limited and focused on public equities.

The survey, conducted from May to July 2025, shows most family offices have minimal exposure to AI through venture capital or growth equity, with over half reporting no such investments.

Despite interest in infrastructure supporting AI, 79% of respondents have no allocations to infrastructure assets.

Family offices are increasingly considering investments in data centers and related services without choosing specific AI winners.

Additionally, 89% of family offices reported no exposure to cryptocurrencies, but many are increasing their holdings of gold, especially those concerned about geopolitics.

🔗 Source: Bloomberg

🧠 Food for thought

Implications, context, and why it matters.

The AI investment paradox may be a mirage for the ultra-wealthy

JPMorgan’s survey finds a gap between what family offices say about AI and how little they allocate to it, with most exposure sitting in public equities. A separate Goldman Sachs survey of 245 family offices reports 33% invest in cryptocurrency and that many use AI as both an investment theme and an operational tool, including internal process work 1.
Goldman’s research points to some family offices taking more risk and chasing new technology trends. It still does not back a claim that 86% already have AI exposure, or that they are shifting beyond public equities into infrastructure and power “picks and shovels” plays, meaning indirect investments that support AI rather than funding specific AI companies 1.
The gap may come from different respondent pools. Goldman’s group leans toward institutional family offices, more formal and professionally managed family investment organizations, with 67% reporting net worth of at least $1 billion 1.

AI’s real-world footprint creates a new class of industrial investments

Rising AI demand is pushing a big spending wave in physical infrastructure, opening deals beyond software.
One hyperscale AI-focused data center, a very large facility built to run computing workloads at scale, can use as much electricity as 100,000 households. Data centers’ share of global electricity demand is forecast to triple over the next decade 2.
That load strains supply since new generation takes years to build, permits move slowly, and suppliers face shortages for parts like turbines 2.
This backdrop can steer capital toward private assets such as power plants or transmission lines, which may diversify portfolios through low correlation versus public equity returns as framed in a J.P. Morgan Asset Management note 2.