White House AI and Crypto Czar David Sacks escalated his public criticism of Anthropic on Sunday, July 27, 2026, accusing the AI safety company of pursuing a regulatory strategy that would, in effect, cripple open-source artificial intelligence development — even as virtually every other major player in the technology industry has aligned behind open-source AI.

Context

Sacks posted a pointed warning on X, stating that the 'entire tech industry (save for Anthropic) has come out in favor of open source AI' and urging the rest of the industry to 'watch these guys like a hawk.' His post catalogued what he described as a pattern of incremental restrictions dressed in neutral language — limiting who can use open-source models, who can contribute to them, how powerful they can be, and whether safety guardrails can be removed.

The thrust of his argument is that these individually framed limits add up to a de facto ban on open-source AI, and that the industry should not be lulled by assurances that 'nobody is trying to ban open source.' Sacks used the term 'gaslighting' to characterise what he sees as Anthropic's public messaging strategy.

Policy Backdrop

The open-source AI debate has been a live fault line in Washington since at least October 2023, when the Biden administration issued an executive order on AI safety and risk mitigation. That order set the stage for ongoing regulatory discussions around model access, compute thresholds, and safety evaluations.

Meta's release of its Llama 2 open-source model in July 2023 sharpened the divide between safety-first organisations — which argue that unrestricted model access raises catastrophic-risk concerns — and a broader Silicon Valley coalition that views open access as essential to innovation, competition, and national competitiveness. Anthropic, the developer of the Claude family of AI models, has consistently positioned itself on the safety-and-oversight end of that spectrum and has been active in policy advocacy in Washington.

Sacks, who joined the Trump administration as AI and Crypto Czar, has been a vocal champion of pro-innovation AI policy and has previously pushed back against what he characterises as regulatory capture by incumbents seeking to lock in their market positions through legislation.

Stakeholders and Impact

The stakes are significant for a wide range of actors. Independent developers, academic researchers, and smaller AI startups depend on open-source models to build products and conduct research without the cost and access barriers of proprietary APIs. Restrictions on who can use or modify such models could concentrate AI capability in the hands of a handful of well-capitalised companies.

For Anthropic, which has raised billions of dollars and competes directly with OpenAI, Google DeepMind, and Meta, the policy environment is commercially consequential. Safety-oriented guardrails that apply to open-source models but not to closed, API-gated systems could tilt the competitive landscape in favour of companies like Anthropic that operate primarily through controlled access. Critics of such regulation, including Sacks, argue this dynamic is not incidental.

The post arrives at a moment when Congressional committees are actively examining AI model access and export controls, meaning Sacks's public framing of the debate could shape how legislators and other executive-branch officials approach forthcoming hearings and draft legislation.

What's Next

Whether Anthropic responds publicly to Sacks's characterisation remains to be seen. The company has not, as of this post, issued a statement addressing his specific claims. Congressional action on AI model access rules is expected to be a priority in the months ahead, and the White House AI office under Sacks is likely to play a central role in shaping the administration's position.

The broader industry coalition backing open-source AI — which Sacks frames as nearly unanimous — will be watching whether that consensus translates into coordinated lobbying pressure or whether safety-oriented voices succeed in carving out regulatory space for restrictions. The outcome will have lasting consequences for who gets to build, use, and benefit from the next generation of AI models.