California's SB 53 and SB 1047 completely exempt companies below a revenue or training-cost threshold ($500M for SB 53), and Anthropic supported SB 53 and was ambivalent on SB 1047.
Unresolvable The public record could not settle it.
Factual: Statements about the present or the past, checked against the record: did they hold up?
The second judge would not resolve this claim; the verdict shown is the first judge's.
Said for an organisation they lead or work for; it counts on both records.
What would count as true or false
Judge as of 2026-08-15. The claim has three parts: (a) SB 53 completely exempts companies below a revenue or training threshold; (b) SB 1047 did the same; (c) Anthropic supported SB 53 and was ambivalent on SB 1047. TRUE: the enacted or final bill texts show that companies below a stated threshold (revenue, training cost, or the training-compute threshold that defines a covered model) have no obligations under the bill at all, and Anthropic's public record shows support for SB 53 and a mixed or ambivalent stance on SB 1047 (for example a letter weighing benefits against costs without full endorsement). MOSTLY TRUE: all of this holds, but with a small detail slip, such as the threshold being defined by compute rather than cost or revenue, or the parenthetical $500M figure being wrong or applying only to a sub-tier. MIXED: one bill fits the 'completely exempt' description but the other places some obligations on companies below its revenue or cost threshold, or Anthropic's stated stances differ materially from what is described. MOSTLY FALSE: both bills have thresholds, but companies below the stated revenue or cost thresholds still carry real obligations, so 'completely exempt' overstates them. FALSE: neither bill has such an exemption threshold, or Anthropic opposed SB 53.
Set before the research began.
Evidence (4)
-
E1 SB 53/RAISE Act Comparison Chart - fpf.org Future of Privacy Forum (Justine Gluck, AI Policy Analyst) · 8 January 2026
“Cost Threshold: Both laws use a $500M+ annual revenue threshold to define “large” frontier developers, narrowing the scope further. Neither bill is scoped to apply to small developers.”
-
E2 SB 53/RAISE Act Comparison Chart - fpf.org Future of Privacy Forum (Justine Gluck, AI Policy Analyst) · 8 January 2026
“Large Frontier Developer: A frontier developer that together with its affiliates collectively had gross revenues in excess of $500 million (Sec. 22757.11 ( j)).”
-
E3 California’s legislature just passed AI bill SB 1047; here’s why some hope the governor won’t sign it | TechCrunch TechCrunch · 30 August 2024
“After several of Anthropic’s suggested amendments were added to SB 1047, CEO Dario Amodei issued a letter saying the bill’s “benefits likely outweigh its costs.” It’s not an endorsement, but it’s a lukewarm signal of support.”
-
E4 California’s legislature just passed AI bill SB 1047; here’s why some hope the governor won’t sign it | TechCrunch TechCrunch · 30 August 2024
“SB 1047’s rules would only apply to the world’s largest AI models: ones that cost at least $100 million and use 10^26 FLOPS (floating point operations, a way of measuring computation) during training.”
Reasoning
E1 and E2 show SB 53's $500M revenue threshold for large frontier developers, E4 shows SB 1047's cost and compute thresholds, and E3 shows a mixed stance on SB 1047. No evidence item documents Anthropic's support for SB 53, which the criteria require for a true or mostly true verdict, so the evidence does not settle part (c) of the claim.
Cite this claim
Dario Amodei, “California's SB 53 and SB 1047 completely exempt companies below a revenue or training-cost threshold ($500M for SB 53), and Anthropic supported SB 53 and was ambivalent on SB 1047.” (said 15 August 2026). Verdict: unresolvable. Reliability API, claim c_9d7271d48fbf, method 0.1.0+c876030c, dataset release v1.1.2. https://epistacklabs.com/claims/c_9d7271d48fbf
This claim is in dataset release v1.1.2, with the same verdict.
Reply to this claim
More factual claims by Dario Amodei
-
Mixed
“Similarly, the “Pacing the Frontier” letter envisions (or at least Anthropic’s preferred implementation of it envisions) modulating the pace of the very best models while not constraining those who…”
-
Mixed
“This contrasts with six months ago when most of the industry was still pushing for preemption of all state regulation and no apparent federal approach either.”
-
Mostly true
“Anthropic actually has been the most lean forward of all the AI companies in working with the U.S. government and working with the U.S. military.”
-
Mostly true
“all we've received is a tweet. We haven't received an actual supply chain desig-- you know, there's -- there's been no actual action by the government.”
-
Mostly true
“The whole timeline has been driven by the Department of War, not by us.”
The whole timeline of the dispute was driven by the Department of War, not Anthropic.
Verdicts are made automatically; how verdicts are made.