White House Closes In on a Review Window for Frontier Model Launches in the US
The US government is close to finalizing, with OpenAI, Anthropic and Google, a voluntary framework that gives federal agencies 30 days to review a frontier AI model before launch. The legal basis is Executive Order 14409, signed on June 2, 2026, and the formal announcement is expected before August 1.
White House Closes In on a Review Window for Frontier Model Launches in the US
The US government is close to finalizing, with OpenAI, Anthropic and Google, a voluntary framework that gives federal agencies a 30 day window to review a frontier AI model before launch, according to a roundup from AI Weekly and legal analysis published by Mondaq and Lexology. The framework's formal announcement is expected before August 1, 2026.
Rooted in Executive Order 14409
The framework's legal basis is Executive Order 14409, signed by President Donald Trump on June 2, 2026, which created a voluntary government pre-release review mechanism for advanced AI models. In the final version now close to being settled with the labs, the voluntary review window dropped from 90 days, set out in an earlier version of the order, to 30 days.
How the review window works
According to Executive Order 14409 and analyses from IAPP and Fenwick, companies developing systems the government classifies as a covered frontier model can give the federal administration up to 30 days to assess a new model before sharing it with trusted partners; broader public release typically follows that partner phase. The mechanism is voluntary, not a binding legal requirement, but OpenAI, Anthropic and Google have already signaled intent to opt in.
A bigger piece: an AI cybersecurity clearinghouse
The same executive order called for setting up an AI cybersecurity clearinghouse, due to become operational by July 2, 2026, responsible for coordinating AI assisted vulnerability scanning, validating findings and managing patch distribution across critical infrastructure sectors such as healthcare, banking and utilities. That piece shows the framework isn't limited to reviewing models before launch, but also tries to use AI to protect infrastructure already in production.
The risk of delay for whoever depends on early access
According to analysis from byteiota, a voluntary 30 day window could delay access to new capabilities, with real competitive consequences for businesses whose differentiation depends on adopting AI features ahead of competitors. For Brazilian agencies and companies that tend to integrate model features as soon as they ship in closed preview in the United States, the practical effect could be one more delay between a feature's announcement and its actual availability outside a lab's circle of trusted partners.
What's clear
Even though voluntary, the White House framework formalizes a pattern the major US AI labs appear willing to accept: government review before a frontier model's public launch has become expected practice, not an exception. For anyone planning a product launch on top of a new frontier model, it's worth treating this review window as one more variable in the timeline, alongside technical delays and regulatory barriers like the one already affecting Grok 4.5 in the European Union.