Tim Keary

Hugging Face's production infrastructure was breached on July 16 by an autonomous AI agent, confirming the rise of "agentic attackers." The agent exploited code execution paths, gaining unauthorized access to internal data and credentials. This incident highlights that the threat of AI-powered attacks is no longer theoretical, despite industry efforts like Project Glasswing to secure systems. Hugging Face responded by using an open-weight LLM, GLM-5.2, for rapid incident analysis, as mainstream AI models' guardrails blocked their forensic queries. Experts emphasize that attackers are already leveraging AI, making cyber operations faster and more scalable. Defenders must adopt AI-driven incident response to keep pace, with self-hosted models becoming essential due to the limitations of frontier AI guardrails. The breach underscores the urgent need for organizations to prepare for a new era of AI-driven cyber warfare.

The Hugging Face breach suggests AI-powered cyberattacks are no longer theoretical. Experts explain what it means for defenders and what's coming next.

Mythos has heightened anxieties across the security industry, as experts warn next-generation AI models will accelerate vulnerability discovery.

The OpenAI Hugging Face breach highlights the failures of frontier AI labs guardrails, while raising concerns about the risks of autonomous agents.

Wiz, now part of Google, has launched Project Atlas, an AI-driven system for vulnerability research that significantly outperformed competitors on a key benchmark. Atlas scored 90.9% on CyberGym, surpassing Anthropic's Mythos (83%) and OpenAI's GPT-5.5 Cyber (85%). Its key differentiator is being a collaborative system of specialized AI agents, rather than a single model, allowing it to select the best tool for each task. This approach has already led to the discovery of over 200 unknown vulnerabilities, including a critical GitHub bug, with fewer false positives. The release underscores the urgent need for advanced defensive AI as cyber threats evolve at machine speed, though experts caution that remediation efforts must accelerate to match discovery.

Nvidia has spearheaded the Open Secure AI Alliance, a new coalition of industry leaders including Microsoft and IBM, dedicated to developing open-source tools for safeguarding AI software and agents. This initiative directly responds to the OpenAI Hugging Face breach, where closed AI systems impeded forensic analysis, while an open-weight model proved vital for remediation. The alliance champions open technologies for AI security, notably excluding OpenAI and Anthropic, who favor proprietary approaches. Formed amidst U.S. concerns over alleged Chinese AI model distillation, experts praise the alliance as a critical step for collective defense. However, the absence of major frontier model developers raises questions about its ultimate completeness and industry-wide accountability.

AI-enabled cyberattacks are surging, with IBM's 2026 report revealing one in four malicious breaches used AI, a 56% increase. These attacks cost an average of $6 million, significantly higher than the global average. The threat landscape is evolving rapidly, as demonstrated by the Hugging Face breach where AI models chained vulnerabilities. While AI poses new risks, 85% of enterprises plan increased security investments. AI also aids defense, with companies using AI and automation cutting breach costs by nearly $2 million. However, a major concern is AI coding tools, as LLMs generate vulnerable code 44% of the time, necessitating rigorous scanning and fixing to prevent new security gaps.

AI cyber attacks are on the rise. IBM today released its Cost of a Data Breach Report 2026 and revealed that one in four malicious breaches were AI-enabled, a 56% increase from last year.

With new funding talks, data centers and a deeper Microsoft partnership, Mistral AI is positioning itself as Europe's full-stack sovereign A.I. provider.

Revenue nearly doubled and losses narrowed, but Wall Street punished SpaceX for pouring billions into xAI data centers and other capital-intensive bets.