65% of S&P 500 companies now mention AI on their earnings calls. That’s not surprising — what BCG found in a report published on August 13, 2026 is that talking about AI on the earnings call and actually capturing value from it are two increasingly different things. Most companies are in the first group. Very few are in the second. And the reason has nothing to do with which model they use.
The risk that doesn’t show up in this quarter’s results
BCG calls this phenomenon “hollowing.” Revenue keeps coming in, market share looks stable, but margin quietly drains away underneath, because generative AI made it trivial to replicate exactly what used to be a company’s edge: specialized knowledge, tailored analysis, operational efficiency. Once any competitor can buy access to the same model and land on the same output, that edge stops being an edge — it becomes table stakes, identical for everyone.
It’s a dangerous risk precisely because it doesn’t show up where leadership usually looks first. Revenue and market share can look healthy for several quarters straight while margin disappears — and by the time it finally shows up on the income statement, the damage is already done.
The tool was never the advantage
The most common reasoning error is treating “using AI” as a competitive differentiator. It isn’t, for the same reason having email stopped being one after the 2000s: once a capability becomes accessible to every competitor at the same price, it stops differentiating — it just raises the minimum bar to keep competing at all. BCG’s report is blunt about this: language models, text generation, automated analysis — all of it is being commoditized at the pace hyperscalers are investing to make it cheaper and more available (BCG itself estimates $750 billion in hyperscaler capex in 2026 alone).
That’s not a reason to stop using AI — it’s a reason to stop expecting mere usage to produce an advantage. Two companies in the same industry, using the same model for the same type of task, land on nearly identical results. If an advantage exists, it has to live somewhere else.
What separates margin winners from everyone else
BCG identifies a pattern in the companies that keep capturing value even as AI itself gets commoditized: they hold some asset that compounds with use, instead of sitting static. The report’s example is Visa, which processes 257 billion transactions a year — every transaction processed makes the network a little harder to replicate, because the data it generates doesn’t exist anywhere else. A competitor can buy the same AI model Visa uses. It can’t buy 257 billion historical transactions in one purchase.
On the other side, the report describes exactly the most common profile of a vulnerable company: expertise easily replicated by a generic model, static data that doesn’t grow with use, operational excellence with no structural moat behind it. It isn’t about company size — it’s about whether what a company does today can be copied by any competitor with access to the same AI, or depends on something specific to that operation that nobody else has.
As João Paulo Batistella, an innovation executive, argues, this is exactly why formalizing process into data matters more than picking which AI to hire: the data specific to an operation — how it makes decisions, where it fails, what actually works for its specific customer — is the one input a competitor can’t buy by plugging in the same generic model. AI amplifies whatever already exists; if what exists is just tacit knowledge sitting in someone’s head, there’s nothing left to amplify that’s exclusive to that company.
Three questions before celebrating this quarter’s numbers
BCG’s logic applies to the S&P 500, but the question behind it applies to a company of any size already using AI day to day:
- What does this company do today that a competitor could replicate by buying access to the same AI model, at the same price?
- Is there any data — customer data, operational data, past decisions — that only exists because this specific company generated it, and that gets more valuable with every additional use?
- Is this quarter’s margin rising because the operation actually improved, or is it just holding steady because nobody has fully commoditized what the company sells yet?
If the answer to the first two is “not sure,” the AI used so far probably just cut cost — it didn’t build any advantage that survives the next competitor who also decides to use it.
Follow Eleva Tecnologia for more analysis on technology applied to business: follow @ElevaTechno on X or @elevatechnologies on Instagram, or learn about the group at elevatec.net/about.
