NYSE joins ICE Markets in deploying Anthropic’s Claude Mythos for cybersecurity

The New York Stock Exchange and its parent company, Intercontinental Exchange, are now using Anthropic’s Claude Mythos Preview to scan their infrastructure for security flaws. The deployment, announced June 3, makes ICE one of the largest financial institutions to join Anthropic’s Project Glasswing, a cybersecurity initiative that pairs advanced AI with critical infrastructure operators to find and fix software vulnerabilities before attackers do.

What Project Glasswing actually does

Anthropic’s Claude Mythos Preview model is deployed across ICE’s platforms, including its exchanges and clearing houses, to identify vulnerabilities that might otherwise sit undetected for years. During its testing phases, Mythos Preview has already identified thousands of critical-severity vulnerabilities across various codebases. Some of those flaws had been lurking for over two decades, including a 27-year-old vulnerability in OpenBSD and a 16-year-old bug in FFmpeg.

Project Glasswing launched in early April 2026 with an initial cohort of roughly 50 organizational partners. In under two months, that number has ballooned to approximately 200 critical-infrastructure entities. Anthropic is backing the initiative with up to $100 million in usage credits for participating organizations, plus $4 million in donations earmarked for open-source security groups.

Why a stock exchange needs AI security

NYSE President Lynn Martin framed the partnership as a move toward “technology-forward innovations” to protect market infrastructure.

ICE isn’t alone in recognizing this. The Project Glasswing partner list reportedly includes firms like AWS, Apple, and Microsoft, all of which have acknowledged the necessity of AI-enhanced tools for defending software infrastructure.

What this means for investors and digital asset markets

Consider what Mythos Preview already demonstrated: finding a 27-year-old vulnerability in OpenBSD. Crypto smart contracts are notoriously difficult to audit, and many DeFi exploits target bugs that have been sitting in code for months or years. An AI model capable of catching ancient flaws in traditional codebases could be transformative when applied to smart contract auditing, where the stakes are just as high and the code review resources are far thinner.

One risk worth flagging: concentration. If a significant portion of critical financial infrastructure relies on the same AI model from the same provider, a flaw in that model becomes its own systemic vulnerability. Anthropic’s Mythos Preview is still in its preview phase, and while identifying thousands of bugs is impressive, the model’s own reliability at scale in production environments remains to be fully validated.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.