Payward, parent of crypto exchange Kraken has joined to Project Glasswing, an initiative run by Anthropic to give select partners access to cutting-edge AI models for safeguarding critical infrastructure from cyber-attacks.
The company has joined other big names in tech and finance, including Amazon Web Services, Apple, Google, Microsoft, and JPMorganChase. Anthropic’s advanced AI model called Claude Mythos 5, can analyze software at scale, identify and prioritize remediation of vulnerabilities for engineering teams, and mitigate potential exploits before malicious actors can take advantage of them
Rolling Out AI-driven Threat Detection Across Kraken’s Systems
Payward confirmed Monday that it intends to deploy Mythos 5 across its internal environments over the coming weeks, folding whatever the AI uncovers directly into its existing security operations. Beyond just protecting its own systems, the company also plans to pass along any vulnerabilities it discovers in third-party open-source software to the relevant project maintainers, extending the benefit of this AI-assisted scanning well beyond Kraken’s own infrastructure.
Payward has joined @AnthropicAI‘s Project Glasswing, and are actively incorporating Claude Mythos 5, Anthropic’s most capable model for finding and fixing software vulnerabilities, into our defensive cybersecurity work. pic.twitter.com/37qOkAIGeQ
— Payward (@Payward) August 17, 2026
This access became possible after the U.S. government cleared Mythos 5 for use by American organizations responsible for operating and defending critical infrastructure. Anthropic originally launched Project Glasswing back in April 2026, and has steadily broadened participation since then, pulling in more organizations from both the technology and finance sectors.
Why AI Cuts Both Ways In The Crypto Security Battle
The timing here isn’t coincidental. Artificial intelligence has increasingly become a double-edged sword for the crypto industry, arming malicious actors with faster ways to locate weaknesses, automate large-scale attacks, and craft far more convincing phishing and social-engineering schemes.
That dynamic has pushed exchanges and custody providers, already prime targets for hackers, into a race to adopt the same technological edge defensively before attackers gain too much of a head start.
How Secure Is Today’s Crypto Infrastructure?
Payward has pointed out that the operational demands facing crypto platforms mirror those of other critical financial systems. Exchanges, custody solutions, and settlement infrastructure all run continuously, creating high-value, always-on targets that never get a break from potential threats.
Arjun Sethi, co-CEO at Payward, framed the stakes bluntly in the announcement, noting that defenders have traditionally faced a losing battle since attackers only need to find a single flaw while defenders must catch every one, every day, without fail. He described frontier AI as the first real tool capable of shifting that imbalance back in favor of defenders.
What Comes Next For Broader Industry Access?
For now, access to Mythos-class cybersecurity tools remains limited to organizations directly involved in operating or defending critical infrastructure, as Anthropic continues building out the safeguards necessary for a wider release. Still, the company has signaled plans to gradually expand availability as those protective measures mature, suggesting more organizations across finance and crypto could eventually gain similar AI-powered defensive capabilities down the line.
Stay informed with the latest trends in Web3, blockchain innovation, and cybersecurity updates at 3verseTV
You need to login in order to Like
Leave a comment