Goldman Sachs’ latest read on Q2 earnings season delivers an uncomfortable finding about the broader market. Across the S&P 500, investors are still paying for the AI infrastructure story, not for any proven AI productivity payoff. That’s a trend spanning every industry, but it carries particular weight for asset management. BCG’s own annual asset management report shows a version of that same divide already playing out inside the industry itself, separating firms that are capturing AI’s value from those still waiting for it.
Starting with the Goldman report, which found just 2% of S&P 500 companies could actually quantify how much AI added to their earnings this quarter, unchanged from Q1. And while AI infrastructure stocks accounted for roughly half of the S&P 500's 31% earnings growth, the median company grew a more modest 14%--a sign AI productivity is not coming through in industry yet (Goldman Sachs Research, "US Weekly Kickstart," Aug. 14, 2026).
Although asset management companies comprise less than 10% of the names in the S&P 500, our industry can’t afford to treat that as someone else's problem. Among the larger classification of financial services firms specifically, asset managers rank behind banks, fintechs, and insurers in AI maturity, according to BCG's scoring of the sector. This puts the asset management industry in the position of being both a laggard adopter and, for structural reasons, one of the industries under the most pressure to innovate now.
An industry that cannot afford to wait for proof
Economics explains the urgency. More than 80% of asset management revenue growth in 2025 came from market appreciation, not manager skill--the traditional engine of the business--while margins have sat flat at roughly 30% for over a decade because costs have grown faster than revenue. Growth is migrating toward retail investors, defined contribution plans, and digital platforms, with an estimated $124 trillion in wealth expected to change generational hands in the U.S. by 2048. The old model of charging more to manage more is not coming back.
Consequently, AI is not merely a productivity nice-to-have. BCG frames AI as central to that shift, with the potential to cut costs 25% to 35% and expand client coverage per relationship manager three-to-fivefold. That’s a rational bet on paper. Yet there remains the problem that the industry has almost no way to prove the bet is working while the infrastructure is being put in place.
AI’s prospective payoff
Today the AI playbook in asset management is a roadmap, not a result. Yet, the theoretical upside, for those firms that push through the uncertainty of early adoption, can be substantial according to BCG. They estimate 55% to 65% more capacity in fund operations with roughly 40% lower costs, not to mention broader research coverage and faster product-to-market cycles.
Two bifurcations converging
What makes this moment sharper than a routine "AI is changing the industry" story is that AI is one of the few levers available to the asset management industry when it is most in need of a reinvention. We see two forces converging.
The first is the overall split Goldman's earnings data already shows across the broader market. This manifests as a widening gap between companies that have captured visible, provable AI value--so far, almost exclusively the infrastructure providers--and companies still funding AI experiments by quietly reallocating existing software and labor budgets, which describes roughly two-thirds of firms surveyed.
The second is internal to asset management. A small number of firms will do what BCG calls “going deep”--redesigning the operating model, embedding governance from the start, and backing a handful of transformative bets rather than scattering pilots across the org chart. The rest will keep optimizing at the margins. And because AI advantage compounds--more automated capacity funds more experimentation, which funds more advantage--the gap between these two groups won't stay linear. Within a few years, it could come to resemble the gap already visible between AI infrastructure stocks and everyone else waiting for the productivity story to arrive.
Where judgment still wins
As baseline analysis gets commoditized, BCG argues competitive edge shifts to judgment--deciding which models to trust, when to override them, and how to translate machine-generated insight into something a client will actually act on. Distribution and relationships, not technology, become "the primary battleground," because clients will choose the firm they trust to exercise judgment when the system is uncertain, not the firm with the most automated workflow. That's consistent with what I've argued in this column before about AI in wealth management more broadly: trust, not capability, is the real gating factor on adoption.
And we all know that trust-gated adoption is, almost by definition, slow to show up in a quarterly print.
The bet asset managers cannot defer
That 2% figure from Goldman is a market-wide gut check--AI's earnings impact remains narrow, unproven, and concentrated in the companies selling the infrastructure rather than the ones supposedly benefiting from it. For most industries, that's a reason to watch and wait. But asset management doesn't have that luxury. With margins flat for over a decade and growth shifting toward a market AI itself will reshape, the firms that wait for Goldman-grade proof before committing capital will be competing, in five years, against firms that didn't wait. And by then, the gap won't be one earnings season wide. It will be a structural chasm.
This article was written by Carrie McCabe from Forbes and was legally licensed through the DiveMarketplace by Industry Dive. Please direct all licensing questions to legal@industrydive.com.