AI agents could reduce corporate profit margins, analyst suggests
9 reported
A reader named Clifford Sosin has proposed that artificial intelligence may be bad for corporate profit margins, according to a post on Marginal Revolution. Sosin argues that many companies profit because customers do not closely monitor them or insource services, citing Bank of America’s $2 trillion in unoptimized deposits and the difficulty of switching enterprise software vendors. He suggests that AI agents, acting as rational sourcing agents, could take rents from toll-takers and redistribute them to consumers. Sosin notes that the AI stack itself may not generate much profit due to commoditization of models and increased competition in GPUs and chip manufacturing. He acknowledges that some companies will win, such as low-cost providers and those with advantaged business models, but his overriding sense is that the equilibrium outcome is lower margins. Sosin adds that if margins decline by 100–200 basis points from roughly 12%, companies may not see much benefit even as consumers are better off.
What’s reported
Clifford Sosin proposed that AI may be bad for corporate profit margins.
Many companies profit because customers cannot be bothered to monitor them or insource services.
Bank of America has $2 trillion of deposits, none of which is optimized.
Most enterprise software vendors could be switched out more often but it is too much of a pain.
AI agents will take rents previously collected by toll-takers and redistribute them to consumers.
The AI stack itself may not make much profit due to commoditization and competition.
Chip manufacturing may remain high-margin but new entrants could make TSMC’s market more competitive.
Sosin’s overriding sense is that the equilibrium outcome is lower margins for companies.
If margins decline 100–200 basis points from roughly 12%, companies may not see much benefit.
Key figures
Clifford Sosin (reader who proposed the idea)
Sources: marginalrevolution.com
