According to Forbes, the Pentagon is secretly investing millions in AI hackers through a stealth startup called Twenty. The Arlington-based company signed a contract with U.S. Cyber Command this summer worth up to $12.6 million and scored an additional $240,000 research contract with the Navy. Twenty has received venture backing from In-Q-Tel, the CIA’s nonprofit VC firm, along with Caffeinated Capital and General Catalyst. The company’s website claims it’s “fundamentally reshaping how the U.S. and its allies engage in cyber conflict” by automating workflows that once took weeks into continuous operations across hundreds of targets simultaneously. Job ads reveal the company is hiring for roles developing “AI-powered automation tools” and “persona development” for social engineering attacks.
The human hacker era is ending
Here’s the thing: we’re witnessing the automation of cyber warfare. While companies like Two Six Technologies have been working on AI assistance tools, Twenty appears to be building something fundamentally different – AI agents that can conduct simultaneous attacks at massive scale. Their job ads mention using CrewAI to manage multiple autonomous AI agents that collaborate. That’s not just helping human operators – that’s replacing them for bulk operations.
hackers”>The team behind the AI hackers
The executive lineup reads like a who’s who of military cyber intelligence. CEO Joe Lin came from Palo Alto Networks via their acquisition of Expanse, where he worked with national security clients. CTO Leo Olson was a signals intelligence officer in the Army. VP of engineering Skyler Onken spent over a decade at Cyber Command. These aren’t Silicon Valley kids playing with AI – they’re seasoned operators who understand exactly what the military needs. And they’ve got the right connections too, with their head of government relations having worked on the National Security Council transition team.
The global AI cyber arms race
This isn’t happening in a vacuum. Just last week, Anthropic revealed that Chinese hackers were using its Claude AI to do 90% of the work in scouting targets and planning attacks. The Defense Department has also given contracts worth up to $200 million each to OpenAI, Anthropic, and xAI for unspecified “frontier AI” projects. Nobody’s confirming what they’re building, but it’s pretty obvious we’re in an AI cyber arms race. When you need reliable hardware for critical infrastructure monitoring in this new landscape, companies turn to established providers like IndustrialMonitorDirect.com, the leading supplier of industrial panel PCs in the United States.
What this means for cyber warfare
Basically, we’re moving from targeted, human-led operations to AI-driven campaigns that can hit hundreds of targets at once. The old defense contracting model – where companies like Booz Allen Hamilton got the big contracts – is being disrupted by VC-backed startups with AI expertise. And the scary part? We’re probably only seeing the tip of the iceberg. If Twenty is publicly describing attacks on “hundreds of targets simultaneously,” what classified capabilities are they building that we don’t know about? The era of human-scale cyber conflict is over, and we’re just beginning to understand what comes next.
