Key Takeaways
- Anthropic’s proposal to extend Mythos AI access to 70 additional organizations has been rejected by White House officials.
- The powerful AI model was withheld from general availability because of its capacity to identify and exploit security flaws in software systems.
- Select technology giants such as Apple, Microsoft, and Nvidia currently utilize Mythos through an exclusive initiative named “Glasswing.”
- Government officials fear that broadening access could strain computing resources needed for federal operations.
- Security was compromised when unapproved individuals gained Mythos access coinciding with Anthropic’s initial expansion announcement.
In early April 2026, Anthropic introduced Mythos as a groundbreaking AI system engineered to identify and leverage weaknesses in essential software infrastructure. According to the company, this technology could uncover security gaps that remained hidden for years, eluding both seasoned security professionals and conventional scanning systems.
Given these powerful capabilities, Anthropic chose to restrict public distribution of Mythos. The company established an exclusive access program dubbed “Glasswing,” granting a customized version of the model to select corporate partners, including Apple, Microsoft, and Nvidia, for strengthening their cybersecurity frameworks.
Anthropic subsequently sought to broaden this arrangement by adding approximately 70 more organizations, potentially expanding the user base to nearly 120 entities. This expansion strategy has now encountered resistance from the current administration.
Administration Cites National Security and Resource Constraints
According to Bloomberg News, a White House representative speaking anonymously revealed that Trump administration officials have conveyed their disapproval of Anthropic’s expansion proposal.
A primary issue centers on computational infrastructure. Government representatives express apprehension that distributing Mythos to a larger corporate network might compromise federal agencies’ capacity to leverage the technology for their own security requirements, the Wall Street Journal reported.
A White House spokesperson informed the Journal that the administration seeks to harmonize technological advancement with security imperatives while guaranteeing responsible AI deployment.
This opposition intensifies existing friction between Anthropic and the Trump administration. In February, President Trump directed federal agencies to discontinue using Anthropic’s products following Defense Secretary Pete Hegseth’s designation of the company as a supply chain security threat.
Anthropic has initiated legal proceedings to contest these restrictions.
White House Dialogue and Security Breach
There appeared to be progress in Anthropic’s relationship with the administration earlier this month. CEO Dario Amodei participated in discussions with government officials at the White House, which Anthropic characterized as “productive.”
However, complications emerged when Bloomberg News disclosed that unauthorized individuals within a closed online community obtained access to Mythos on the identical day Anthropic publicly revealed its controlled distribution strategy. The company has acknowledged the breach and launched an internal investigation.
Several industry observers have questioned whether Anthropic has exaggerated Mythos’s performance capabilities. The company faces intense competition, particularly from OpenAI, which is developing comparable cybersecurity AI tools.
Anthropic has not issued a statement regarding the administration’s opposition. White House representatives did not provide comments when contacted.
