Summary
CNBC reports on the widening rift between Anthropic and the Trump administration, centering on AI czar David Sacks’s accusation that Anthropic is running a “regulatory capture strategy based on fear-mongering.” While OpenAI has embedded itself as a White House partner (Stargate, White House dinners), Anthropic has consistently opposed the administration’s AI regulatory positions and endorsed California’s transparency requirements. The piece illuminates how AI companies are splitting along political lines.
Key Points
- David Sacks accused Anthropic of “regulatory capture through fear-mongering” after Jack Clark’s “Technological Optimism and Appropriate Fear” essay
- OpenAI aligned with Trump from day one: Stargate joint venture announced Jan 21, 2025
- Anthropic blocked Trump’s attempt to preempt state AI regulation for 10 years (the “Big Beautiful Bill” provision, ultimately abandoned)
- Anthropic endorsed California SB 53 requiring transparency and safety disclosures — directly opposite to administration’s approach
- OpenAI valuation: $500B; Anthropic: $183B
- Anthropic holds a $200M DoD contract and GSA App Store access despite conflict
- Anthropic hired senior Biden-era officials for its government relations team
- Dario Amodei compared Trump to a “feudal warlord” during 2024 election; publicly supported Kamala Harris
- Amodei was not invited to White House dinners attended by Meta, OpenAI, and Nvidia leaders
- Anthropic offered a version of Claude to government customers for $1/year (showing it wants federal business while opposing policy)
Newsletter Angles
- The AI industry’s political split: OpenAI chose access, Anthropic chose principles — and is paying a price. This is a real-time case study in how tech companies navigate state cooptation
- The “regulatory capture” accusation is itself a rhetorical move: Sacks is accusing Anthropic of exactly what all tech companies do — shaping regulation to favor incumbents. The difference is direction
- Anthropic’s contradictory position: actively seeking federal contracts at $1/year while opposing the administration. The commercial necessity of government as a customer may ultimately force alignment
- Jack Clark’s essay is the flashpoint — worth reading as a standalone artifact (see Import AI 431 source page)
Entities Mentioned
- Anthropic — central subject; the company opposing Trump’s AI regulatory agenda
- OpenAI — contrasting subject; aligned with White House, Stargate partner
- Dario Amodei — Anthropic CEO; compared Trump to feudal warlord, supported Harris
- Tech-State Conflict — the structural dynamic this story exemplifies
Concepts Mentioned
- Tech-State Conflict — Anthropic is the clearest current case of a tech company in direct ideological conflict with the executive branch
- Regulatory Weaponization — Sacks’s “fear-mongering” framing attempts to delegitimize Anthropic’s safety arguments
Quotes
“Anthropic is running a sophisticated regulatory capture strategy based on fear-mongering.” — David Sacks
“SB 53’s transparency requirements will have an important impact on frontier AI safety. Without it, labs with increasingly powerful models could face growing incentives to dial back their own safety and disclosure programs in order to compete.” — Anthropic blog post
“If Anthropic actually believed their rhetoric about safety, they can always shut down the company. And lobby then.” — Keith Rabois
Notes
Published October 2025. Complements the earlier Britain/DoD conflict story (which covers 2026 events). This source captures the political split before the DoD blacklisting — it’s the precursor phase.