Fast Facts

• The US is investing millions in AI-driven cyber offense, led by stealthy startups like Twenty.
• AI agents can simultaneously scan, infiltrate, and attack hundreds of targets with minimal human input.
• Venture capital from intelligence-linked funds, including In-Q-Tel, is fueling rapid development.
• China is reportedly running similar automated cyber operations, escalating the arms race.
• Automated systems blur the line between traditional hacking and mass-scale digital warfare.

The Dawn of Automated Cyber Offense

Picture a war room with no human operators, just silent rows of servers, each running tireless digital agents. In this new era, the United States is rewriting the rules of cyber conflict, deploying AI agents that can probe, plan, and penetrate enemy networks at a scale and speed never seen before. The concept, dubbed “multi-threaded hacking,” means attacks happen in parallel, not one by one - like unleashing a swarm instead of a single spy.

Meet the New Players: Startups and Spies

At the heart of this revolution is Twenty, an Arlington-based startup operating mostly under the radar. With contracts totaling over $12 million from US Cyber Command and the Navy, Twenty is not your typical defense contractor. Its backers include In-Q-Tel, the venture arm tied to the CIA, and other high-risk tech investors. What sets Twenty apart is its mission: transforming labor-intensive cyber operations into automated, continuous pipelines that can target vast digital territory simultaneously.

Job postings at Twenty offer a glimpse inside: engineers are tasked with building frameworks that map out attack paths, coordinate fleets of AI agents, and even create digital personas for social engineering - tricking adversaries into revealing secrets or opening doors, not with brute force, but with cunning conversation.

The Global Cyber Arms Race

The US is not alone. Recent research from Anthropic shows Chinese cyber groups are already leveraging AI models to automate reconnaissance and attack planning, dramatically shrinking the time between identifying a target and launching an assault. The Pentagon, meanwhile, has inked multi-million dollar deals with AI giants like OpenAI and Anthropic, though details remain top secret. This shadowy competition is accelerating the pace and scale of cyber operations worldwide, pushing digital conflict toward automation and mass effect.

Other companies, like Two Six Technologies, are developing autonomous modules capable of launching attacks when success is likely, and Israeli firm Tenzai is using AI for defensive simulations. But Twenty's approach - massively parallel, automated offense - signals a new phase: warfare not just faster, but fundamentally transformed by AI.

Redefining Digital Battlefields

In the past, skilled human hackers painstakingly mapped networks and crafted exploits. Now, AI agents can scan, plan, and strike hundreds of targets at once, turning cyber offense into an assembly line. It’s as if the lone pickpocket has been replaced by an army of invisible, tireless burglars, all acting at digital speed. As the US pours resources into this new frontier, the line between defense and offense blurs, and the stakes for global cyber stability rise.

WIKICROOK

• AI Agent: An AI agent is an autonomous software program that uses artificial intelligence to perform tasks or make decisions for users or systems.
• Multi: Multi refers to using a combination of different technologies or systems - like LEO and GEO satellites - to improve reliability, coverage, and security.
• Social Engineering: Social engineering is the use of deception by hackers to trick people into revealing confidential information or providing unauthorized system access.
• Reconnaissance: Reconnaissance is the early stage of a cyberattack where attackers gather information about a target to identify weaknesses and plan their approach.
• Automation Pipeline: An automation pipeline is a system that connects and automates tasks, enabling processes to run efficiently with minimal human intervention.