The 'regulatory capture' proxy fight on record was lab against lab, both of them incumbents.
White House AI czar David Sacks repeated his claim that Dario Amodei's safety advocacy is 'regulatory capture' and 'a DMV for AI,' the same week Politico framed Anthropic's PAC spending as an 'anti-AI blitz' and the All-In circle amplified both.
Safety advocacy that also serves a well-capitalized lab's interests is how most industry policy positions work, and the spending is real. The specific charge, rules designed to exempt Anthropic and crush smaller rivals, needs evidence the record doesn't provide.
The one concrete electoral fight ran frontier lab against frontier lab: Anthropic-aligned money countering OpenAI-aligned money over a state safety-disclosure candidate. And Amodei's own stated design goal runs the opposite way, though that's his account, unaudited here.
Two checks: on whether Anthropic's own proposals shield incumbents against smaller rivals, and on what the one documented electoral proxy fight actually pitted against what.
Being right about what happened and wrong about what it proves are different failures, so these two lines never get averaged into one score. How this works
The logic check
If the claim is right, these things should be true. Here's what we found. Each check says which version of the claim it's testing, the plain one most people would hear or the widest one the same words could carry, and whether it carries the claim's own burden or establishes the background it sits on.
If Anthropic's regulatory push is designed to entrench itself and disadvantage smaller or open-source competitors, Anthropic's own account of its policy proposals should show them shielding or advantaging large frontier incumbents like itself relative to smaller rivals.
Reading strainedAmodei states publicly that Anthropic 'tries very hard to make proposals that disadvantage (slow down) frontier AI companies while advantaging smaller competitors,' the opposite of an incumbent-protecting design, and explicitly rejects the 'regulation = regulatory capture = concentration of power' shorthand as 'an overly simplified picture of the world.' source
If Anthropic's spending is specifically self-interested capture protecting its own frontier position against rivals, the one documented electoral proxy fight its money engaged in should show it defending rules that carve out exemptions for large incumbents while burdening smaller AI startups, rather than a fight against a rival frontier lab's own political spending.
Reading failsNPR's reporting on the closest documented electoral proxy fight, the NY-12 Bores primary, shows Anthropic-linked super PAC money as part of over $23 million in combined pro- and anti- spending countering OpenAI-aligned super PAC money that was itself trying to defeat a candidate backing state safety-disclosure legislation; the fight was frontier lab against frontier lab, Anthropic-aligned spending against OpenAI-aligned spending, not incumbent against startup. source
The full scoring
The raw numbers behind the verdict. What R/I/P/E/L and charge stages mean: how we score.
Research verdict: mixed evidence. The research found some of what the claim predicts and not the rest.
The claim, from David Sacks and the All-In circle, is that Anthropic's political spending and safety advocacy amount to regulatory capture: rules only Anthropic can meet, entrenching it while wiping out startup and open-source rivals. Amodei says Anthropic 'tries very hard to make proposals that disadvantage (slow down) frontier AI companies while advantaging smaller competitors,' but that's his word without an audit of the bill text, so that test lands mixed. The one documented electoral fight, the NY-12 Bores primary, shows Anthropic-linked super PAC money countering OpenAI-aligned money in over $23 million of combined spending, frontier lab against frontier lab, with rival labs' money on both sides. The spending is real and the startup-crushing design isn't in the record.
Source record: hobocode.net Fault Line Report 2026-W34
The Fault Line Report, August 17 to 23, 2026
All eleven narratives with the full logic test, the mirror check, the R/I/P/E/L factors and the coverage notes on each lane. This page is the plain-language cut; that one is the whole thing, including the polarization methodology that defines every score here.