Anthropic Claude Mythos Preview Sparks Wall Street Panic: Bessent, Powell Summon CEOs; Cloudflare Tumbles 8%
AI Podcast
U.S. Treasury Secretary and Federal Reserve Chair met with Wall Street bank CEOs regarding cybersecurity risks from Anthropic's AI model, Mythos. This highlights AI's growing threat to financial systems, especially due to legacy code in core banking infrastructure. Anthropic launched Project Glasswing with tech giants to scan for vulnerabilities, but cybersecurity stocks declined, fearing disruption by AI-driven remediation. The utilities sector also faces risks from potentially slower-to-update critical infrastructure.
TradingKey - According to the latest Bloomberg report, U.S. Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell urgently convened a meeting with the CEOs of major Wall Street banks this Tuesday to discuss potential cybersecurity threats posed by Anthropic's latest AI model, Mythos.
Analysts pointed out that the emergency meeting sends a significant signal: the rapidly advancing capability of AI models to identify and exploit vulnerabilities has become one of the primary risks currently facing the financial industry.
According to sources familiar with the matter, the meeting's participants included the CEOs of nearly all major U.S. systemically important banks, including Citigroup's (C) Jane Fraser, Morgan Stanley's (MS) Ted Pick, Bank of America's (BAC) Brian Moynihan, Wells Fargo's (WFC) Charlie Scharf, and Goldman Sachs' (GS) David Solomon, all of whom were in attendance.
The reason Mythos is so dangerous is that the model can, under user instructions, "identify and exploit vulnerabilities in every mainstream operating system and web browser," Anthropic stated.
Anthropic stated that this model is too powerful, so there are currently no plans to open it to the public. Furthermore, on April 7, Anthropic announced the launch of an industry joint initiative named "Project Glasswing," partnering with Amazon (AMZN) , Apple (AAPL) , Microsoft (MSFT) , Cisco (CSCO) and other companies to provide tools for using the Claude Mythos Preview model for vulnerability scanning and remediation of critical software infrastructure, aiming to reinforce the security of key systems before similar AI models become widely prevalent.
Mythos Threatens Vulnerable US Banking Core Systems
JPMorgan Chase (JPM) JPMorgan Chase Chief Information Security Officer Pat Opet stated that the bank will evaluate the tool's application value in defending critical financial infrastructure in a "rigorous and independent manner."
The reason the financial sector may face cybersecurity impacts is that many banks' core systems run on legacy code, which creates inherent security risks.
For example, the Hogan system used in JPMorgan's retail banking core dates back to the 1980s and 1990s, and it was only after a 2021 partnership with fintech firm Thought Machine that the bank planned a gradual replacement. After decades of operation, the system has accumulated countless layers of patches, making it difficult to update or repair, thereby leaving openings for AI models to exploit vulnerabilities. Furthermore, core systems at banks like Citigroup and HSBC have also been running for decades.
Cloudflare Plummets 8%; Utility Infrastructure at Risk
In February this year, Anthropic released new security features for its Claude AI model capable of scanning codebases for security vulnerabilities, triggering a massive sell-off in the cybersecurity sector, and the current release of Mythos is no exception. At the close of U.S. markets on April 9, cybersecurity stocks fell across the board, with Cloudflare (NET) dropping 8.62%, while CrowdStrike (CRWD) fell 7.46%.
It is worth noting that Cloudflare, which had previously benefited from Anthropic's rise, slumped this time because of it. The reason is that cybersecurity stocks, represented by Cloudflare, have not only been excluded from Anthropic's AI beneficiary circle but may even suffer significant harm.
Anthropic co-founded Project Glasswing to study system security issues, but Cloudflare was excluded, leaving it without first-hand information for strengthening its systems. Furthermore, if large models like Mythos can autonomously defend and remediate vulnerabilities in real-time, what business value remains for providers like Cloudflare that offer traditional reactive monitoring and protection layers?
Beyond cybersecurity stocks, another potential victim is the utilities sector, which is typically viewed as a defensive play during the AI wave but now faces risks of its own. This is because core system software updates for the power grid, water resources, and natural gas pipelines may be slower than those in banking, with even longer life cycles; once system security is compromised, large-scale outages or physical damage could result in more tangible losses.
This content was translated using AI and reviewed for clarity. It is for informational purposes only.
Recommended Articles
