Anthropic Chief Executive Officer Dario Amodei finds himself at odds with fellow technology leaders over the pace of artificial intelligence development and the necessity of strict new regulations.
The friction centers on warnings Amodei published in a September 12, 2026, essay calling for a deliberate slowdown in advanced AI capabilities to allow safety research and independent evaluation to catch up.
The disagreement surfaced publicly at a White House event on September 29, 2026, where at least two dozen artificial intelligence executives signed voluntary safety principles. Both Amodei and Nvidia Chief Executive Officer Jensen Huang were among six corporate signers of the White House Accord on Super Intelligence finalized that same day.
The voluntary agreement calls for internal controls, an internal oversight team, independent external evaluations, and board oversight, while also requiring regular meetings among signers on safety standards and noting that controls could later be written into law or regulation.
Despite the shared signatures, executives hold sharply opposing views on whether government intervention is required. Huang and other industry leaders have pushed back against Amodei’s calls for regulation, arguing that market-driven solutions and internal corporate controls are sufficient. In private meetings following the White House event, Nvidia’s chief executive described Amodei’s warnings as overly extreme.
There’s no conflict between innovation, technology, and safety.
Jensen Huang, Nvidia CEO
Earlier in July, Microsoft Chief Executive Satya Nadella and Jensen Huang had also intervened when leading AI companies lobbied the White House to prevent certain models from being developed and shared.
The Microsoft and Nvidia executives emailed each other drafts of a letter stating that those AI models should remain open, quickly gathering support from two dozen companies before the letter became public, followed by 200 more companies—including OpenAI—signing on within the next week.
In contrast, executives from companies like Nvidia and Microsoft have viewed those same powerful models primarily as major business opportunities rather than imminent existential hazards.
While Amodei has urged ongoing access for independent evaluators, coordination among frontier AI companies in democratic countries on safety standards and development pace, and cooperation between democratic and authoritarian governments, Anthropic has absorbed tangible financial and operational expenses for its cautious approach. The company’s internal safeguards and emphasis on restraint have directly resulted in delayed model releases.
Earlier coverage: Investors Question Anthropic's $1.5 Trillion IPO Valuation
