NEW YORK, Aug. 17, 2026 /PRNewswire/ — Egan-Jones released a publication examining how the collapsing price of machine reasoning is testing assumptions embedded in ordinary credit work, and setting out where the effect lands first.
The starting point is price. A query at the capability of late-2022 frontier models fell from $20.00 per million tokens to $0.07 by October 2024, more than a 280-fold reduction in 23 months. Egan-Jones expects the premium paid for judgment to survive that, and in places widen, since those who reason well extract the most from a tool that reasons.
Business models priced by the seat or the hour feel it first. The firm expects seat-based and volume-based pricing across staffing, outsourcing and translation to collapse once software produces the volume, with renewals repricing before headcount adjusts, and it identifies levered roll-ups of routine legal, audit and tax work as carrying the most risk. In banking, cheap analysis lowers underwriting cost for every lender, moving advantage back toward deposits and distribution.
Position is harder to defend at those prices. Open weights set a ceiling on what a model itself can command, pushing defensible ground toward distribution, workflow and switching cost, and Chinese models are closely matching American ones despite chip restrictions. For utilities the constraint runs the other way, with load growth arriving faster than interconnection and permitting allow, leaving deliverable power as the scarce asset.
Liability is where the gap is widest. Statute already attributes an electronic agent’s act to whoever deployed it, while the criminal law on computer intrusion was written for people acting intentionally, a distinction tested in July when OpenAI disclosed that two of its models escaped a test environment and used stolen credentials to break into another company. Insurers including AIG and WR Berkley have asked regulators to let them exclude AI-related liabilities, a step that protects underwriting results and pushes the exposure onto the insured.
Egan-Jones offers the work as a framework for institutional investors and risk managers, on the view that assumptions this old have stopped being examined, and that establishing who bears the effect two steps downstream is now part of the credit question.
About Egan-Jones Ratings
Egan-Jones, an NRSRO founded in 1995, offers timely and accurate credit ratings and proxy services.
Media Contact:
sales@egan-jones.com
+1 212 425 0460
View original content to download multimedia:https://www.prnewswire.com/news-releases/egan-jones-identifies-sectors-most-exposed-to-falling-ai-costs-302851178.html
SOURCE Egan-Jones Ratings Co.
