Artificial intelligence developer Anthropic announced on Tuesday a significant expansion of Project Glasswing, a collaborative industry effort designed to identify and patch critical software flaws using advanced machine learning. The initiative is now extending its reach to approximately 150 additional organizations distributed across more than 15 nations.

This operational milestone arrives shortly after the technology firm initiated a confidential filing for an initial public offering. That financial move closely follows a massive funding phase that secured $65 billion, propelling the company to a valuation approaching $1 trillion.

Core Technology and Sector Expansion

The driving force behind Project Glasswing is Claude Mythos, an artificial intelligence model that the developer describes as its most robust creation to date. The system has demonstrated the capacity to detect thousands of previously unknown zero-day vulnerabilities within a matter of weeks. During an initial deployment phase in early April, the firm granted 50 foundational partners, notably including federal departments within the United States government, early access to the Claude Mythos Preview to evaluate their internal code repositories.

With the latest expansion, the program is incorporating sectors that were not heavily featured in the preliminary rollout. The newly added participants operate within critical infrastructure domains such as power generation, water treatment, healthcare services, telecommunications, and hardware manufacturing. According to corporate statements, a significant portion of these new participants includes commercial enterprises and nonprofit entities that maintain foundational software relied upon by global governments and other businesses.

The firm highlighted the high stakes involved, noting that the shared characteristic among these partners is the severe consequence of a potential security breach. Company representatives estimated that a substantial compromise in these foundational systems could impact upward of 100 million individuals, carrying profound implications for both national and international security.

International Participants and Market Competition

The roster of newly integrated organizations spans multiple countries that maintain strong diplomatic ties with the United States. Based on reports from financial media outlets referencing sources close to the initiative, the geographic footprint includes Australia, Canada, France, Germany, Italy, Switzerland, the Netherlands, Spain, Belgium, Sweden, India, Japan, New Zealand, and South Korea.

These reports also identified several specific entities that have reportedly received access to the Mythos system. The list features American identity management provider Okta, alongside major South Korean technology conglomerates Samsung, SK Hynix, and SK Telecom. Additionally, multinational organizations such as NATO and the European Union Agency for Cybersecurity (ENISA) are said to be participating. Official confirmation regarding these specific partnerships remains pending.

The urgency to establish robust security protocols through Project Glasswing is driven by the rapid pace of artificial intelligence development. Anthropic executives have publicly projected that competing technology firms will soon release models matching the capabilities of the Mythos Preview.

Validating these industry expectations, competitor OpenAI recently introduced its own security-oriented model, designated as GPT-5.5-Cyber. That system has already been deployed to a substantial network of external partners for initial testing and evaluation.