Kraken’s parent company, Payward, has joined Anthropic’s Project Glasswing, a program built to provide vetted organizations with access to advanced cybersecurity AI. The move places Payward in a collaborative framework designed to bolster security research by leveraging Anthropic’s AI models to hunt for vulnerabilities in software and systems, according to Decrypt. The arrangement is described as part of a broader effort to strengthen defenses around crypto exchange operations by incorporating cutting-edge AI-assisted security assessments.
Decrypt notes that Project Glasswing is Anthropic’s initiative to extend access to its cybersecurity-focused AI to select institutions and groups that meet specific vetting criteria. In this context, Payward will gain access to Anthropic’s capabilities, including the model known as Claude Mythos, which is targeted at identifying security weaknesses and potential exploits. The integration aims to enable more thorough probing of security architectures and to facilitate a structured approach to discovering and reporting flaws, the report adds.
CoinDesk corroborates the development, highlighting that Payward’s involvement with Glasswing centers on security testing and vulnerability hunting. The report explains that Payward intends to use the cybersecurity model to systematically search for vulnerabilities and to share open-source findings, contributing to a broader ecosystem of security research and transparency. This collaboration aligns with ongoing industry efforts to improve the resilience of digital-asset platforms by encouraging rigorous, machine-assisted security reviews rather than relying solely on internal testing.
The collaboration comes amid a continuing focus on security within the crypto exchange landscape, where operators face constant scrutiny over potential weaknesses that could be exploited by attackers. By granting Payward access to Claude Mythos through Glasswing, Anthropic aims to enable more rapid and comprehensive analysis of security postures. The program’s vetted participants are expected to conduct ongoing tests, document vulnerabilities, and disseminate non-sensitive findings to the wider community, helping other platforms shore up defenses without exposing proprietary details.
Both outlets describe the move as part of a larger trend of crypto-native firms engaging with AI-powered security tools and third-party research networks. The articles emphasize that the initiative is not about deploying AI to operate exchanges or automate decision-making in trading but about enhancing defensive capabilities. The emphasis remains on responsible disclosure, controlled access, and collaboration with researchers to ensure that any identified flaws are addressed promptly and that information sharing remains structured and safe for users.
Industry observers noted that the adoption of AI-assisted security testing could influence how exchanges and related platforms allocate resources for cybersecurity. While the reports do not specify timelines, the partnership signals a commitment to rigorous, ongoing security evaluation that leverages external perspectives and advanced AI tools. For investors and users, the development underscores a growing priority on safeguarding digital assets and reducing the likelihood of breaches that could undermine user trust and market stability.
In summary, Payward’s participation in Anthropic’s Glasswing program marks a notable step for Kraken’s corporate security philosophy. By accessing Claude Mythos to hunt vulnerabilities and publish open-source findings, the company aligns with broader industry practices that favor collaborative, AI-enabled security research. The reports from Decrypt and CoinDesk together depict a partnership framed around enhanced defensive capabilities, transparency, and shared learning within the crypto ecosystem.