#590
AI at $1.8 Trillion Scale

Norway’s $1.8t oil fund makes more than 46m equity trades a year — and it’s using AI to make fewer of them. By predicting index changes, internalizing trades across teams, and better timing execution, the world’s largest sovereign wealth fund expects to cut $400m in annual trading costs. That’s not a backtest — it’s already saved $100m.

The Norwegian fund is among the largest equity investors in the world, owning an average 1.5% of every listed company globally. By handling more trades internally — and better predicting when to execute or hold — it seeks to reduce both transaction costs and market impact. And it’s not alone: nine out of ten investment managers surveyed by Mercer last year said they already use or plan to use AI, with two-thirds citing cost reduction as a primary goal.

These results from production-grade AI aren’t unique to asset managers. UnitedHealth Group now runs over 1,000 live AI applications — half powered by generative models — across insurance, delivery, and pharmacy units. These systems support everything from transcribing clinician visits and and guiding patient referrals to tracking down missing claims data — embedding AI into the company's everyday workflows.

These real-world deployments signal a shift: industry incumbents aren’t merely experimenting with AI — they’re weaving it into their core operations. A sovereign wealth fund trims basis points; a healthcare giant slashes paperwork. This isn’t about flashy demos anymore — it’s about large-scale, operational deployment by institutions already embedded in the fabric of the global economy.