Hitachi accelerates critical infrastructure security validation processes with Claude Mythos Preview through Anthropic’s Project Glasswing

• Participated in Anthropic-led Project Glasswing and conducted verification utilizing Claude Mythos Preview to enhance the cybersecurity of Hitachi’s software and products for social infrastructure 
• Significantly reduced the verification process, including threat modeling for source code bases comprising millions of lines of code, from several weeks, demonstrating Mythos’ advanced analytical capabilities and confirming its applicability to real-world operations 
• Will apply insights gained to the development and operation of software and products for social infrastructure to help mitigate cyberattack risks against critical infrastructure systems

Tokyo, July 28, 2026 Hitachi, Ltd. (TSE:6501, "Hitachi"), is advancing efforts to identify and remediate vulnerabilities using Claude Mythos Preview (“Mythos”), an advanced AI model, through Project Glasswing, an AI-powered security initiative led by Anthropic PBC (“Anthropic”), which Hitachi announced it joined in June 2026^*1.

Hitachi strongly supports the initiative’s mission to help protect the world’s most critical software and advance cybersecurity practices required in the age of increasingly capable AI. Through the application of Mythos across more than 100 use cases within the Hitachi Group, Hitachi identified potentially severe vulnerabilities and confirmed the effectiveness of strengthening security throughout the development and operational processes of its software and products supporting critical infrastructure.

Specifically, Hitachi confirmed that threat modeling for source code bases comprising millions of lines of code can be completed within a few hours using Mythos—a task that previously required security specialists and developers several weeks to complete. In addition, while conventional AI-based vulnerability detection often identifies potential vulnerabilities that require substantial human effort to verify false positives, Mythos autonomously generates proof-of-concept exploit procedures and test code for runtime validation, enabling verification of findings at the time of scanning. As a result, Hitachi confirmed improved vulnerability detection accuracy and a significant reduction in developer validation effort. Based on these findings, Hitachi will continue to implement timely remediation measures and product updates to help reduce cyberattack risks to critical infrastructure systems.

• 1 Hitachi Press Release dated June 5, 2026: Hitachi joins Anthropic’s “Project Glasswing” an AI-powered security program

Comment from Hiroaki Koiwa, Head of Cyber CoE, AI & Software Service Business Unit, Hitachi, Ltd.

“As AI continues to evolve, increasingly sophisticated cyberattacks are becoming easier to execute, making stronger security measures more important than ever for critical infrastructure. Through this validation initiative, we have become convinced of the transformative potential that frontier AI can bring to cybersecurity. While AI can rapidly perform large-scale software analysis, human expertise remains essential for final assessments and risk decisions. The combination of AI capabilities and cybersecurity professionals’ expertise will be a cornerstone of next-generation cyber defense. Going forward, we will leverage the knowledge gained through this initiative not only to strengthen the security of software and products supporting social infrastructure, but also to further enhance HMAX by Hitachi (“HMAX”), a suite of next-generation solutions that brings the power of AI to social infrastructure. Through these efforts, we will contribute to the realization of safer, more resilient social infrastructure.”

Key Outcomes of the Project Glasswing Validation Initiative

(1) Identification and remediation of advanced vulnerabilities in Hitachi’s software and products

Using Mythos, Hitachi identified potential vulnerabilities in its mission-critical software and publicly exposed APIs comprising millions of lines of code that had not been detected by existing security tools. The initiative went beyond simple bug discovery and established an end-to-end process, from detailed triage based on AI-generated proof-of-concept code to evaluation of concrete remediation measures. The results demonstrated a highly effective approach to proactively improving the quality and security of products that support critical infrastructure.

(2) Validation of frontier AI’s advanced reasoning capabilities

For threat modeling and large-scale source code analysis, Hitachi evaluated a unique approach that combines Mythos’ exceptional contextual understanding with autonomous AI agents capable of iterative reasoning and experimentation, together with a specialized “harness” framework that codifies the thought processes of cybersecurity experts. This approach enabled assessments at a scale and speed unattainable by human experts alone while producing highly accurate and practical results.

(3) Acquisition of knowledge and best practices for deployment in development and operational processes

Through this initiative, Hitachi gained valuable insights into AI-assisted vulnerability assessment methodologies, evaluation and triage processes for AI-generated findings, effective collaboration between AI and human experts, and considerations for applying these technologies to software used in critical infrastructure environments. Going forward, Hitachi will leverage these insights to promote a “shift-left” security approach that embeds security from the earliest stages of development and incorporate them into the development and operational processes of HMAX and other social infrastructure solutions.

Related Links

About Hitachi, Ltd.

Through its Social Innovation Business (SIB) that brings together IT, OT(Operational Technology) and products, Hitachi aims to be a global leader in continuously transforming social infrastructure through digital, contributing to a harmonized society where the environment, wellbeing, and economic growth are in balance. 

Hitachi operates worldwide across four sectors – Digital Systems & Services, Energy, Mobility, and Connective Industries – as well as a Strategic SIB Business Unit focused on new growth areas. With Lumada at its core, Hitachi creates value by combining data, technology and domain knowledge to solve customer and social challenges. 

Revenues for FY2025 (ended March 31, 2026) totaled 10,586.7 billion yen, with 606 consolidated subsidiaries and approximately 290,000 employees worldwide. Visit us at www.hitachi.com.