Khaberni - In a move that reflects the escalating global cyber arms race, the U.S. War Department, formerly known as "Defense", has this year enhanced its offensive capabilities in cyberspace, by contracting with a highly secretive startup called "Twenty", which is working on developing artificial intelligence agents capable of executing large-scale, automated hacking operations against foreign targets considered by Washington to be a threat to its interests.
According to federal contract records, the company, based in Arlington, Virginia, has received a contract with the U.S. Cyber Command worth $12.6 million, along with a research contract worth $240,000 with the U.S. Navy. These contracts are rare cases where a venture-backed startup has obtained direct offensive projects with U.S. government cybersecurity entities, as reported by Forbes.
"Twenty" is supported by influential entities in the intelligence and security investment community, notably In-Q-Tel, the investment arm associated with the Central Intelligence Agency (CIA), as well as investment funds like Caffeinated Capital and General Catalyst. The company has not issued any comments on the contracts or the nature of its work, as it aims to maintain a high level of secrecy.
Although the company has not yet launched its services to the public, its website indicates the development of systems capable of transforming operations that used to take weeks of human work into automated tasks that are continuously and synchronously executed across hundreds of targets.
"Twenty" describes itself as contributing to "reshaping the way the United States and its allies conduct cyber conflicts".
Job advertisements show that the company is developing advanced offensive capabilities including automation tools powered by artificial intelligence and frameworks for attack pathways, and it also relies on open-source platforms such as CrewAI which manages independent artificial intelligence agents capable of collaborating in task execution.
It also refers to its work in developing "digital personas", a method used in social engineering operations to create fake digital identities to infiltrate opponents' networks.
The company is led by an executive team including former military and intelligence agency personnel, among them the CEO Joe Lin, a former officer in the Navy Reserve and former director at Palo Alto Networks, alongside officials who previously worked in the military and the U.S. Cyber Command.
This American move comes at a time when Anthropic revealed last week that Chinese hacking groups used its tools to carry out hacking operations through artificial intelligence agents who perform the majority of targeting and analysis tasks.
With the escalation of the global cyber polarization, it appears that the entry of artificial intelligence into offensive attacks is evolving at a faster pace than expected, heralding a new phase of conflict in which countries rely more on automated systems to carry out digital operations across borders.
