Verizon has joined Anthropic’s Project Glasswing to evaluate advanced AI models for identifying and mitigating vulnerabilities in critical infrastructure, including telecommunications networks. The initiative gives a select group of security organizations access to Anthropic’s Claude Mythos Preview model, which Anthropic says can identify complex software and operating system vulnerabilities at unprecedented speed.
Verizon said it is the only telecommunications provider participating in the program. The company’s information security team has spent recent months testing Claude Mythos Preview to assess how the technology could strengthen network protection and accelerate vulnerability analysis across Verizon’s infrastructure. The initiative also emphasizes cross-industry collaboration, with participating organizations expected to share findings and operational best practices related to AI-driven cybersecurity.
Project Glasswing reflects a broader industry push to apply large language models and agentic AI systems to cybersecurity operations, threat detection, and infrastructure resilience. Telecom operators increasingly view AI-assisted security analysis as critical as networks expand to support AI workloads, edge computing, cloud-native architectures, and highly distributed infrastructure. Verizon said all testing involving Mythos Preview follows strict internal governance and safety standards for evaluating emerging technologies.
- Verizon is participating in Anthropic’s Project Glasswing cybersecurity initiative.
- The project focuses on securing critical infrastructure in the AI era.
- Verizon is testing Anthropic’s Claude Mythos Preview model for vulnerability detection.
- Anthropic said the model can identify complex software and operating system vulnerabilities at very high speed.
- Verizon said it is the only telecom operator currently participating in the initiative.
- The program encourages participating organizations to share cybersecurity insights and best practices.
- Verizon stated that all testing follows rigorous internal safety and governance policies.
“Our customers rely on the security of our network every day. As part of Project Glasswing, we are able to test and improve our cybersecurity efforts with new insights to maintain our network’s security,” said Dan Schulman, CEO of Verizon. “Over the past several months, our information security team has been rigorously testing this critical new technology to determine its benefits to our network. As the only telecommunications company utilizing Mythos Preview, we are uniquely positioned to share cross-industry insights that will help secure the global internet fabric and support our mission to deliver a safe and reliable experience for every customer.”
🌐 Analysis: Anthropic’s decision to tightly control access to Claude Mythos Preview highlights growing concern around dual-use AI models capable of advanced vulnerability discovery. Rather than broadly releasing the capability, Anthropic appears to be following a curated deployment model similar to how hyperscalers and governments handle sensitive cyber tooling. Verizon’s participation gives the telecom sector direct exposure to emerging AI-assisted offensive and defensive security techniques that could reshape network operations and infrastructure assurance.
Anthropic has also referenced participation from “over 40 organizations” spanning technology, cybersecurity, infrastructure, and open-source ecosystems, but many have not been publicly named yet. Known participants in Anthropic’s Project Glasswing include: Intel, Amazon Web Services, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorgan Chase, Linux Foundation, Microsoft, NVIDIA, and Palo Alto Networks.
🌐 Analysis: Telecom operators face mounting pressure to automate cybersecurity workflows as networks become more software-defined, cloud-native, and AI-centric. Verizon has steadily expanded its focus on AI-driven network operations, security analytics, and autonomous infrastructure management. At the same time, competitors and hyperscalers are integrating generative AI into SOC operations, threat hunting, and infrastructure monitoring, signaling a broader convergence between AI infrastructure and telecom cybersecurity operations.
