Are companies benefiting from AI adoption?

AI adoption and monetization are distinct concepts, with only adoption happening at scale in Corporate America. Roughly two-thirds of S&P 500 companies discuss AI on earnings calls, but only 17% have offered specific evidence from its usage, and just 2% have quantified its impact on earnings.¹

So far, the clearest financial winners have been companies selling AI infrastructure. Semiconductor stocks have climbed more than 90% this year, versus roughly 15% for the S&P 500, while AI infrastructure beneficiaries are expected to deliver more than half of the index's earnings growth this season. Evidence is emerging, though, that companies outside tech are also benefiting.

The measurement problem

AI adoption is broad but shallow. Nearly two-thirds of large companies are building or deploying AI agents, yet barely one in ten has scaled them across the business.² Even where deployed, translating productivity gains into revenue growth or margin expansion isn’t straightforward.

Time saved doesn't necessarily mean money saved. American Express, for instance, has cut coding cycle times by more than 30%, but is using the extra capacity to tackle a project backlog rather than reduce engineering costs.³ Netflix is producing higher-quality content faster and more cheaply using AI, but plans to reinvest the savings into more content.⁴ These are real productivity gains, even without visible margin or earnings impact. Meanwhile, AI initiatives often run alongside broader restructuring and automation, making their individual contribution difficult to isolate. Some benefits, like improved quality or customer retention, are harder to quantify altogether.

What separates the early leaders?

Successful long-term investing requires identifying which companies are gaining a durable competitive advantage. Four characteristics may help distinguish early leaders in AI:

1. Proprietary data
   As powerful models become widely available, the differentiator may become proprietary data, particularly information competitors cannot replicate. For instance, Waste Management’s Smart Truck platform draws on proprietary route and service data, generating more than $300 million in annual run-rate EBITDA.⁵
2. Domain and regulatory expertise
   In highly regulated industries, deploying AI requires near perfect accuracy. Outputs must be reliable, explainable and auditable, potentially favoring incumbents with deep industry knowledge and established processes. UnitedHealth's AI-driven prior authorization, one of health care’s most scrutinized processes, reports a 96% first-pass approval rate.⁶
3. A system of record
   Companies holding authoritative records for transactions, credit files or financial benchmarks have a distinct advantage. Many are deeply embedded in customers' workflows, positioning them well to integrate AI into existing services. As workflows become agentic, that combination of trusted data and established customer relationships will become increasingly valuable.⁷
4. Network effects
   AI gives networks another way to become more valuable: Data from one customer can improve the product for everyone else. In insurance analytics, insurers contribute claims data to a shared system, and as more participants join, the models access a wider range of information, potentially improving their usefulness across the network.

Picking winners is still challenging

These characteristics may offer an edge, but do not guarantee success. AI will likely reshape industries, business models and customer behavior. The pace of corporate adoption, though, suggests this transformation will unfold more gradually than frontier technological advances might imply. Today's advantages may not be permanent, but could persist long enough for companies to capitalize on them.

For investors, that creates opportunities among less obvious AI beneficiaries who are converting AI adoption into monetization through these durable advantages. Identifying them will require selectivity, while diversification across the AI value chain remains essential for portfolios.

30a677f4-c716-11f1-975a-6b6d65810aba