US Officials Warn Wall Street of AI Cybersecurity Risks from Anthropic's Mythos Model

• Anthropic and OpenAI are both developing advanced AI cybersecurity products in a competitive race to control offensive and defensive capabilities, with OpenAI finalizing a security product for limited partner release and Anthropic running Project Glasswing internally. • Project Glasswing brings together major tech companies including Amazon, Apple, Microsoft, Google, and others to proactively hunt critical software vulnerabilities before attackers exploit them. • A joint study by Anthropic and MATS Fellows found that Claude Sonnet and GPT-5 could produce exploits against Ethereum smart contracts worth $4.6 million and uncovered two novel zero-day vulnerabilities in nearly 3,000 recently deployed contracts.

• French cybersecurity funding has undergone a structural shift driven by AI-powered threats, regulatory requirements, and U.S. expansion, with Q1 2026 alone surpassing all of 2025's funding activity. • The sector is emerging as a frontline investment priority, though a government report warns of a €3 billion funding gap over five years for breakthrough technologies needed for future defense. • The acceleration reflects growing global demand for advanced security solutions as AI capabilities reshape both offensive and defensive cybersecurity strategies.

• OpenAI is finalizing an advanced cybersecurity product designed to defend critical systems, planning to release it first to a limited set of partners as reported by Axios. • The move represents a strategic shift for frontier AI firms to demonstrate their ability to support defensive security infrastructure amid growing concerns about offensive AI misuse. • AI companies face mounting pressure to prove their models can protect critical systems rather than enable attackers, reshaping how they frame commercial opportunities in the security sector.

• Anthropic announced Project Glasswing, a collaborative initiative involving Amazon Web Services, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, the Linux Foundation, Microsoft, NVIDIA, and Palo Alto Networks to secure critical software infrastructure. • The initiative provides select firms access to Claude Mythos2 preview model for defensive cybersecurity work, which has reportedly identified thousands of severe vulnerabilities in major operating systems and browsers. • The program aims to give defenders a durable advantage in the AI-driven cybersecurity era while controlling access to prevent misuse by attackers.

• OpenAI testified in support of an Illinois bill that shields AI labs from liability for products causing 'critical harm,' including potential mass deaths or catastrophic financial losses. • The legislation is part of a broader industry push amid ramped-up state and federal AI enforcement, such as Florida's probe into OpenAI for potential criminal harm. • Similar liability-limiting measures are under consideration in at least three other states, according to policy trackers monitoring AI legislation.

• Nava, an AI infrastructure startup, raised $22 million in funding led by Greenoaks Capital to expand its GPU cloud and data center capabilities. • The investment targets growing US demand for high-performance computing to train and deploy advanced AI models. • Nava's platform supports AI developers with scalable resources amid chip shortages and cloud competition.

• Anthropic launched Claude Mythos, a cybersecurity AI model that has discovered thousands of zero-day bugs, enhancing defenses against exploits. • The preview version identifies vulnerabilities in open-source projects and will soon charge for detailed release information to clients. • While boosting cybersecurity, the tool raises concerns over potential misuse by threat actors to weaponize the findings.

• Dropzone AI, a Seattle-based startup, secured $16.8 million in funding to develop autonomous AI agents for cybersecurity operations. • The company is a finalist for GeekWire Awards' AI Innovation of the Year, competing with firms transforming HR, retail, biotech, and more. • Investors back the platform amid surging demand for AI-driven threat detection and response in US enterprises.

• Zencoder (For Good AI Inc.) unveiled Zenflow Work, expanding its AI orchestration platform to automate planning, coordination, and engineering in business units like product, marketing, sales, finance, and HR. • The platform addresses tasks beyond coding agents, targeting everyday business workflows. • This launch positions Zencoder as a comprehensive AI solution for enterprise efficiency in US tech sectors.

• Anthropic launched Project Glasswing, collaborating with Apple, Google, and Microsoft to deploy Claude Mythos Preview in identifying and patching critical software vulnerabilities preemptively. • The initiative unites rivals to leverage AI against cyber threats targeting US software infrastructure. • By fixing flaws before exploitation, the project aims to bolster national cybersecurity amid escalating AI-driven attacks.

• Amazon Web Services (AWS) achieved a $15 billion annual run rate in AI revenue during Q1 2026, according to CEO Andy Jassy's annual shareholder letter. • Amazon's internal chips business now generates over $20 billion per year, supporting AI infrastructure and custom silicon development. • The figures underscore Amazon's aggressive push into AI and cloud computing dominance amid intensifying competition from Microsoft Azure and Google Cloud.