Anthropic, the California-based artificial intelligence company best known for its Claude chatbot, has expanded access to its powerful Mythos AI model to approximately 150 organisations across more than 15 countries, including several non-English-speaking nations. The expansion, part of a programme called Project Glasswing, marks a significant broadening of what began as a tightly controlled pilot involving around 50 mostly American technology firms.

Mythos attracted global attention and concern when early testing showed it was exceptionally capable at identifying vulnerabilities in software code — uncovering thousands of security flaws in short periods of time. Anthropic initially restricted access in early April after those tests, fearing the model could be weaponised by hackers if released too broadly. Early participants included major technology companies such as Amazon, Google, Nvidia, Apple and Microsoft, as well as cybersecurity specialists CrowdStrike and Palo Alto Networks. By late May, those partners had already used Mythos to identify more than 10,000 serious security weaknesses.

The new wave of participants extends into sectors that were absent from the original group, including power grids, water systems, hospitals, telephone networks and hardware manufacturers — infrastructure that millions of people and governments rely on daily. Anthropic has warned that a successful cyberattack on any one of these organisations could affect more than 100 million people, with significant consequences for both national and global security. The European Commission confirmed that Anthropic had invited ENISA, the European Union's cybersecurity agency, to join the programme. German security officials have previously cautioned that criminal groups or state actors could attempt to exploit Mythos for offensive cyberattacks. Anthropic declined to name the new partners.

The expansion comes alongside other significant corporate developments. On Monday, Anthropic announced it had filed confidentially for an initial public offering (IPO) on US markets — a move that could make it one of the largest stock market debuts of the year. The company recently closed a funding round valuing it at around $900 billion. Separately, Anthropic is engaged in a legal dispute with the US Department of Defense after the company refused to allow unrestricted military use of its AI technology, stating it would not permit its systems to be used for mass domestic surveillance or fully autonomous weapons.

The expansion of Project Glasswing also reflects shifting attitudes within the US government. Despite an earlier legal dispute, federal authorities eventually agreed to test Mythos themselves over security concerns — a development some observers connect to a broader policy shift under the Trump administration, which has sought agreements with leading AI companies to have their most powerful models reviewed for safety before commercial release. A presidential order formalising that policy was drafted but remains on hold following internal disagreements.