The Pentagon Is Spending Millions On AI Hackers

Key Concepts

• AI in Cyber Warfare: The increasing use of Artificial Intelligence for both offensive and defensive cyber operations.
• AI Agents: Autonomous AI systems capable of performing tasks, specifically in this context, cyber attacks.
• Offensive Cyber Capabilities: Tools and techniques used to disrupt, damage, or gain unauthorized access to enemy systems.
• Social Engineering: Manipulating individuals to divulge confidential information.
• Persona Development: Creating convincing fake online identities for deceptive purposes.
• Automation in Cyber Conflict: Utilizing AI to automate tasks previously requiring significant manual effort.
• Red Teaming: A security practice involving authorized simulated attacks to identify vulnerabilities.

---

Pentagon Investment in Offensive AI Cyber Capabilities

The US Department of Defense is significantly investing in Artificial Intelligence (AI) for offensive cyber warfare, demonstrated by recent contracts awarded to emerging startups. This investment signals a shift towards greater automation and scale in cyber operations. Specifically, a Virginia-based startup, 20 or XX, secured a $6.6 million contract with US Cyber Command in the summer of 2025, alongside a $240,000 Navy research contract. This is notable as such contracts typically favor established defense contractors like Booz Allen Hamilton or L3Harris. 20 received Venture Capital (VC) funding from InQoutell Caffeinated Capital, General Catalyst, and a CIA-founded nonprofit.

20’s Capabilities and Focus

20’s website states its core mission is “Transforming workflows that once took weeks of manual effort into automated continuous operations across hundreds of targets simultaneously,” indicating a focus on large-scale, automated cyber attacks. The company positions itself as “fundamentally reshaping how the US and its allies engage in cyber conflict.” Job postings reveal the development of “advanced offensive cyber capabilities,” including “attack path frameworks” and “AI powered automation tools.” They are utilizing open-source tools like Creive, designed for managing collaborative autonomous AI agents. A key aspect of their strategy involves “persona development,” suggesting the creation of convincing fake online accounts for social engineering and potential government cyber attacks.

Leadership and Expertise

20’s executive team comprises individuals with extensive backgrounds in military and intelligence. CEO and co-founder Joe Lynn is a former US Navy Reserve officer and previously held a VP role at PaloAlto Networks, following their acquisition of Expanse, where he focused on national security client vulnerability assessments. CTO Leo Olsen also has experience with Expanse’s national security team and served as a signals intelligence officer in the US Army. Skyler Enen, VP of Engineering, spent over a decade at US Cyber Command and the US Army. Adam Howard, Head of Government Relations, previously worked on the National Security Council transition team for the Trump administration.

Global Landscape of AI in Cyber Warfare

The US is not alone in leveraging AI for cyber warfare. Anthropic reported in November that Chinese hackers automated 90% of their cyber attack processes using its tools, encompassing target scouting and attack strategy. The US Defense Department has awarded contracts worth up to $200 million to OpenAI, Anthropic, and XAI (Elon Musk’s AI company) for “frontier AI projects,” though the specific applications remain undisclosed.

Comparison with Existing AI Cyber Tools

While 20’s approach appears to emphasize large-scale automation, other companies are also developing AI-powered cyber tools. 26 Technologies received a $190 million contract by 2024 for its AI offensive cyber tools, primarily designed to assist human operators. Their six-year project, “Ike,” aims to automate support for cyber battle space and strategy development, potentially launching attacks based on probability of success. However, there is currently no evidence that Ike operates at the scale of autonomous AI agent deployment claimed by 20.

Defensive Applications of AI

AI is more commonly deployed for defensive cybersecurity purposes. Israeli startup Tenzi is utilizing AI models from OpenAI and Anthropic to identify vulnerabilities in customer software through “red teaming” exercises – authorized simulated attacks to improve security. This differs from the offensive focus of companies like 20 and 26 Technologies.

Technical Terms Explained

• Creive: An open-source tool for managing collaborative autonomous AI agents.
• Attack Path Frameworks: Systems for identifying and exploiting potential routes to compromise a target network.
• Signals Intelligence (SIGINT): Gathering intelligence from electronic signals and systems.
• Red Teaming: A security exercise where a team attempts to penetrate a system to identify vulnerabilities.
• Frontier AI: Cutting-edge AI technology with significant potential capabilities.

Logical Connections

The report establishes a clear progression from the Pentagon’s investment in AI cyber capabilities to the specific case of 20, detailing its capabilities, leadership, and potential impact. It then contextualizes this development within the broader global landscape of AI in cyber warfare, comparing 20’s approach to existing tools and highlighting the prevalence of AI in defensive cybersecurity.

Data and Statistics

• $6.6 million: Contract awarded to 20 by US Cyber Command (Summer 2025).
• $240,000: Navy research contract awarded to 20.
• 90%: Percentage of Chinese cyber attacks automated using Anthropic’s tools.
• $200 million: Maximum contract value awarded to OpenAI, Anthropic, and XAI by the US Defense Department.
• $190 million: Contract awarded to 26 Technologies by 2024.

Conclusion

The US is actively investing in AI-driven offensive cyber capabilities, exemplified by the secretive startup 20. This investment represents a significant shift towards automation and scale in cyber warfare, potentially reshaping the future of conflict. While AI is more commonly used for defensive purposes, the increasing adoption of AI agents for offensive operations by both the US and its adversaries underscores the growing importance of this technology in the realm of cybersecurity and national security.