TLDR: DARPA has announced Team Atlanta as the $4 million grand prize winner of its two-year AI Cyber Challenge (AIxCC) at DEF CON 33. The competition aimed to develop advanced artificial intelligence systems capable of autonomously identifying and patching software vulnerabilities, marking a significant leap in automated code security.
LAS VEGAS – The U.S. Defense Advanced Research Projects Agency (DARPA) concluded its two-year AI Cyber Challenge (AIxCC) at DEF CON 33 on August 8, 2025, announcing ‘Team Atlanta’ as the grand prize winner of $4 million. This groundbreaking competition sought to push the boundaries of artificial intelligence in automatically finding and fixing software vulnerabilities, a critical need in an escalating cybersecurity landscape.
Team Atlanta, a formidable consortium of tech experts from Georgia Tech, Samsung Research, the Korea Advanced Institute of Science & Technology (KAIST), and the Pohang University of Science and Technology (POSTECH), demonstrated superior capability in the challenge. According to DARPA officials, Team Atlanta “performed best at finding and proving vulnerabilities, generating patches, pairing vulnerabilities and patches, and scoring with the highest rate of accurate and quality submissions.”
The AIxCC, which commenced in 2023, pitted dozens of teams against each other to develop AI-enabled software, known as cyber reasoning systems (CRSs), that could autonomously secure critical code. The competition’s final round presented teams with a daunting task: to analyze 54 million lines of code and identify and patch synthetic vulnerabilities within a four-hour cloud compute time limit. The finalists collectively discovered 54 out of 70 synthetic vulnerabilities (a 77% success rate) and patched 43 (61%). Remarkably, they also uncovered 18 real-world software vulnerabilities, which are now being disclosed to developers.
Andrew Carney, AIxCC Program Manager, lauded the participants, stating, “The world is different because the AI Cyber Challenge (AIxCC) has fundamentally changed our understanding of what is possible in terms of automatically finding, but more importantly, fixing vulnerabilities in software.” He further emphasized, “There is no excuse not to leverage this flavor of automation. And it will only get better. This is the new floor.”
Trail of Bits, a New York City-based cybersecurity firm, secured second place, receiving $3 million, while Theori, comprising AI researchers and security professionals from the U.S. and South Korea, took third place with a $1.5 million prize. Taesoo Kim, a professor at Georgia Tech and leader of Team Atlanta, revealed that his team plans to donate a significant portion of their winnings back to Georgia Tech to further their research, envisioning a future where AI agents serve as de-facto security experts for developers.
Also Read:
- Black Hat USA 2025: Cybersecurity Innovations Highlight AI-Driven Defenses and Emerging Threats
- ADVERSA AI Secures Gold Stevie® Award for Groundbreaking Agentic AI Security Platform
The initiative saw significant collaboration, with the Defense Department partnering with the Health and Human Services Department (HHS) and leading AI companies such as Anthropic, Google, and OpenAI, which provided technical support and $350,000 in large language model credits. Microsoft and the Linux Foundation’s Open Source Security Foundation also contributed assistance. DARPA Director Stephen Winchell announced that four of the seven cyber reasoning systems developed by the finalists would be immediately released as open-source tools for cyber defenders, with the remaining three to follow in the coming weeks, aiming to foster a new ecosystem of autonomous vulnerability remediation.
