Finding big money for AI and a smaller world for security at Black Hat USA 2026
Las Vegas was hotter than hell last week, but not as hot as the market for artificial intelligence-enabled security at Black Hat USA 2026. A bandwagon of million-dollar booths for overfunded agentic security startups arose like mirages from the desert. Fortunately, there was also real value to be found amidst the AI hallucinations. AI supply […] The post Finding big money for AI and a smaller…
Black Hat USA 2026 drew in a flood of AI-enabled security booths, with million-dollar investments pouring into overfunded agentic startups. Amidst the hype, valuable insights emerged about AI's role in security. Chief information security officers and researchers grappled with AI supply chain challenges, such as Anthropic PBC Mythos-class exploits and recent attacks from OpenAI, Hugging Face, and even Meta Platforms Inc. Patrick Duffy, Dropzone AI's head of product, highlighted how AI's nondeterministic capabilities could lead to more code and traffic generated rapidly, making it harder for human teams to investigate.
Despite AI automation not replacing professional security expertise, AI agents could help security analysts by taking on repetitive tasks, freeing up time for critical investigations. However, there are many ways to implement these agents correctly. Nebulock Inc. offers a "hunt-first" approach using a behavioral world-model graph, while Huntress Labs provides a managed platform with in-context training and SOC workflow automation.
Strike 48 offers a comprehensive agentic SecOps platform with connectivity across multiple platforms, and Zero Networks emphasizes the importance of setting boundaries for AI agents to prevent them from becoming adversaries. Geordie AI helps organizations adopt agents securely by discovering, learning, and monitoring them. With AI agents becoming more autonomous, ensuring they act with intent and purpose is crucial for maintaining security.
Written by urgent.news from SiliconANGLE's reporting — not their text. Machine-written — it may contain errors, so check the original before relying on it.
