Breaking

NYSE Used Anthropic’s Project Glasswing to Find and Fix Cybersecurity Vulnerabilities

September 3, 2026 at 01:47 EDT

Security
Foundation Models
Sales & Business

The New York Stock Exchange used Anthropic’s Project Glasswing to identify and address vulnerabilities in its cybersecurity systems, NYSE President Lynn Martin told Congress on Sept. 2. Martin said the exchange was an early participant and had quickly addressed a variety of issues identified through the program. Bloomberg Law

SEPT. 2, 2026 · CRITICAL INFRASTRUCTURE

NYSE puts frontier AI to work on cyber flaws

Project Glasswing gave the exchange access to Claude Mythos Preview to identify and address vulnerabilities in its systems.

The reported scale of discovery

10,000+

high- or critical-severity vulnerabilities across partners

2,000

bugs found across Cloudflare critical-path systems

271

Firefox 150 vulnerabilities found and fixed

Cloudflare findings

The 400 high- or critical-severity bugs are part of the 2,000 total.

400

2,000

High or criticalAll bugs

Defensive upside

NYSE says the technology helped it find issues that could be addressed quickly. Cloudflare and Mozilla also reported substantial findings and fixes.

The constraint

Discovery is accelerating faster than verification, disclosure and broad patch deployment, while public access remains restricted.

A controlled-access model for now

Anthropic has not released Mythos-class models publicly, citing the risk that similar capabilities could make software exploitation cheaper and faster.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.