Verizon has become the only telecommunications company participating in Anthropic’s Project Glasswing, an exclusive initiative that grants access to the AI company’s most advanced security model for protecting critical infrastructure. The partnership allows Verizon to test Claude Mythos Preview, a specialized AI model capable of identifying complex vulnerabilities in software and operating systems at exceptional speed.
The collaboration positions Verizon at the center of a broader industry effort to strengthen cybersecurity defenses as AI capabilities rapidly advance. Project Glasswing brings together select global security leaders who are committed to sharing findings and best practices across industries, creating a collective defense network for essential infrastructure.
Project Glasswing emerged after Anthropic announced that Claude Mythos Preview could detect sophisticated vulnerabilities at unprecedented speeds. Given the model’s powerful capabilities, Anthropic chose to limit access to a carefully selected group of organizations responsible for protecting critical infrastructure rather than releasing it broadly.
“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, Verizon CEO. “Over the past several months, our information security team has been rigorously testing this critical new technology to determine its benefits to our network.”
Verizon’s participation carries significant implications for the telecommunications industry. As the backbone of digital communications, telecom networks represent prime targets for cyberattacks that could disrupt everything from emergency services to financial systems. The company’s unique position as the sole carrier in the program means it will likely influence how AI-powered security tools are deployed across the broader telecommunications sector.
The timing of this partnership reflects growing concerns about cybersecurity threats targeting critical infrastructure. Recent years have seen increasing attacks on power grids, water systems, and communication networks, prompting both government agencies and private companies to seek more sophisticated defense mechanisms. AI-powered security tools like Claude Mythos Preview represent a new frontier in this arms race between attackers and defenders.
Verizon emphasizes that its testing follows strict safety protocols. The company applies rigorous safety standards to every phase of the Mythos Preview evaluation process, ensuring that the powerful AI tool is used responsibly. This cautious approach reflects broader industry awareness about the potential risks of advanced AI systems, even when they’re designed for defensive purposes.
The collaboration also highlights the competitive advantages that early access to advanced AI tools can provide. While Verizon gains insights that could strengthen its network security, the company is also positioned to influence the development of future AI security tools through its feedback and testing results. This could translate into long-term benefits for both Verizon’s network resilience and its competitive position in the telecommunications market.
As Project Glasswing participants share their findings across industries, the initiative could accelerate the adoption of AI-powered security tools beyond telecommunications. The cross-industry knowledge sharing component means that insights gained from protecting telecom networks could benefit other critical infrastructure sectors, from energy to transportation systems.