Trump calls AI oversight a ‘globalist scheme’ as Amodei and Altman head to the UN to ask for it

President Trump rejected international AI oversight at the UN, while OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei are scheduled to urge greater international cooperation and AI-safety measures before the UN Security Council. More than 20 primarily European countries are seeking binding rules for advanced AI, but the U.S. and China have abstained, underscoring a widening policy split between the two leading AI powers and Europe. The administration has already applied pressure to Anthropic, including a Pentagon supply-chain-risk designation and an alleged delay to its Mythos model release, creating incremental regulatory and operational uncertainty for frontier AI companies.

Analysis

The investable implication is not a near-term UN-policy outcome but a widening split between commercial AI deployment and national-security procurement. A permissive domestic posture lowers the probability of broad frontier-model compliance costs, supporting capex intensity for NVDA and hyperscalers; yet it simultaneously raises the value of firms able to satisfy classified-work, auditability, and secure-deployment requirements. That favors GOOG and MSFT/AMZN more than pure frontier labs if defense agencies increasingly treat model access as a supply-chain issue.

NVDA is not a clean regulatory winner: lighter U.S. model rules sustain accelerator demand, but AI negotiations with China elevate the probability that chip-access restrictions become the bargaining instrument. The asymmetry is unfavorable over the next 1-3 months because a new export-control headline can impair the China revenue narrative immediately, while incremental domestic deregulation mainly protects already-embedded demand. AAPL has modest direct upside from a looser framework, but its higher-value outcome depends on whether regulatory fragmentation makes privacy-preserving, on-device AI a differentiated product feature over the next 6-18 months.

Consensus may overstate the relevance of multilateral safety rhetoric to public equities. Binding international rules remain unlikely without U.S.-China alignment; the more likely structural outcome is competing AI governance blocs, raising compliance and localization costs for global platforms while reinforcing the strategic value of U.S. compute infrastructure. The thesis fails if federal enforcement shifts from rhetoric toward model-release restrictions, or if new China controls materially reduce NVDA data-center guidance or hyperscaler capex commitments.

Key Decisions for Investors

• Maintain a core NVDA long, but hedge the next 1-3 months of policy-event risk with put spreads dated beyond the next earnings release; treat any confirmed tightening of China accelerator rules or a reduction in China-related guidance as the stop signal rather than adding into the first headline-driven decline.
• Prefer GOOG over AAPL for a 6-18 month AI-regulatory dispersion trade: Google has greater enterprise, cloud, and security monetization optionality if compliance becomes segmented by jurisdiction, whereas AAPL's benefit remains contingent on consumer adoption of on-device AI features. Reassess on Cloud growth, AI-related capex, and any material safety-driven product delay.
• Watch AMZN and MSFT as second-order beneficiaries of a procurement shift toward auditable sovereign-cloud AI. Do not initiate solely on the UN session; upgrade only if agency contract awards, classified-cloud demand, or management commentary demonstrate that security requirements are diverting workloads from less trusted model providers.
• Use a long NVDA / short broad software basket hedge rather than a directional software short if AI-security incidents drive enterprise caution: infrastructure spend can remain resilient while application-layer deployment cycles lengthen. Exit the hedge if enterprise AI bookings and remaining performance obligations accelerate across the next reporting cycle.