IBM announced the latest expansion of its enterprise security program for the AI era, and is partnering with Anthropic as a member of Project Glasswing, an industry initiative to defend the world’s critical software infrastructure.
IBM is widening its AI-powered security portfolio as the market moves from simple threat detection toward AI-assisted vulnerability research and defense.
Anthropic’s Mythos Preview can identify vulnerabilities across major operating systems and web browsers, and the model found “thousands” of significant flaws. That is the kind of capability enterprise security teams now have to plan around, especially in sectors with legacy systems and complex software stacks.
AI-powered attacks have already moved beyond what traditional defenses can match. We’re helping clients assess their exposure and putting tools like IBM Concert to work in more environments. Separately, as part of Project Glasswing, we’ve been hardening our own products and contributing fixes back to the open-source community. The collaboration makes the entire ecosystem stronger,” said Rob Thomas, SVP Software & Chief Commercial Officer, IBM.
[Also Read: IBM and Aramco Explore AI Collaboration to Advance Industrial Innovation in Saudi Arabia ]
Project Glasswing and the threat picture
Project Glasswing is important because it shows how fast the cyber landscape is changing. Anthropic initially limited the program, then later allowed participants to share findings, best practices and code with outside organizations, regulators and the media under responsible disclosure norms. The Bank of Spain also cited Glasswing as one of the protective AI tools that could strengthen cyber resilience, while warning that advanced models like Mythos could accelerate the exploitation of vulnerabilities and make attacks more synchronized.
That matters for IBM because enterprise buyers now want two things at once: better defense and tighter control. They want AI systems that can scan for weaknesses, but they also want those systems wrapped in policy, audit trails and governance. Reuters’ reporting on Anthropic shows why that demand is rising. When a model can spot vulnerabilities quickly, the time between discovery and exploitation can shrink sharply.
[Also Read: Anthropic Launches ‘Project Glasswing’ with Tech Giants to Boost Global Cybersecurity ]
Security teams will increasingly need tools that can test code, expose weak points and support incident response before attackers do. The market is moving toward AI systems that work on both sides of the security equation, and IBM is clearly trying to stay on the defensive side of that line.
For the wider industry, this is another sign that frontier AI is forcing security vendors, cloud providers and model developers into closer coordination. Project Glasswing is not just a research exercise. It is a preview of how cyber defense may be organized around advanced models, controlled access and rapid disclosure in the years ahead.
