White House Intervenes in OpenAI, Anthropic Releases as AI Oversight Tightens

The White House has intervened directly in the release and export arrangements for advanced models from OpenAI and Anthropic, citing national security. The move signals that AI safety reviews are shifting from corporate self-regulation toward government control. Whether models can be launched and which customers receive priority access could increasingly depend on White House approval and export rules.

The latest developments show that the Trump administration has partially lifted restrictions on Anthropic’s Mythos 5 while retaining export controls on advanced models. OpenAI’s GPT-5.6 Sol may be supplied only to 20 “trusted partners” approved by the White House. As of July 19, 2026, the available information did not disclose the exact date of the orders or cite any transaction value.