Armor has launched Sovereign AI, a governed AI work platform for regulated industries that provides private AI use, company owned data controls, spend management, and audit trails.
Harness has integrated its agentic security capabilities with Google Cloud's Apigee API management platform. The integration helps joint customers discover, monitor, and protect AI agents, AI services, tools, and MCP workflows.
IDMEXPRESS launched IDMEXPRESS.AI with two enterprise security platforms, PulseAI for AI agent security and NHIDM for non human identity management.
Bold Security has announced a new endpoint data protection layer for AI interactions, covering prompts, clipboard activity, file access, MCP payloads, and commands issued by agents.
Invicti Security has launched Invicti Agentic Pentest, a web application and API penetration testing capability that combines autonomous AI reasoning with DAST validation.
Check Point Software Technologies Ltd. has launched the Check Point AI Network Firewall as part of firewall software release R82.20. The product adds AI traffic controls to existing Check Point firewalls and is available now.
Cantina has launched from stealth with $8 million in new funding led by Framework Ventures, bringing its total raised to $16.5 million.
Discern Security raised $13 million in Series A funding led by Forgepoint Capital and launched Discern Security Loop, an agentic platform for finding and fixing security control gaps.
Reco has expanded AI Runtime with Browser Guard, prompt and tool call analysis, and Smart Remediation. The new capabilities add browser visibility, runtime inspection, and action recommendations for AI agent security findings.
FAR.AI launched an AI Security Leaderboard comparing safeguard resistance across frontier models in CBRNE and cybersecurity misuse tests. Its first results found hundreds of universal jailbreaks in Grok 4.5 and Gemini 3.1 Pro, and none in Claude Fable 5 or GPT-5.6 Sol.
Defense Unicorns and Paramify have partnered to integrate Paramify's compliance automation platform into the Unified Defense Stack for federal and defense customers.
Pentera has added AI native web application testing to its security validation platform, with beta access for selected customers and general availability planned for Q4 2026.
Anthropic says Claude Mythos Preview found new mathematical attacks against HAWK and a seven round AES research variant. The company says the results do not affect current production encryption systems.
The FCC has selected Quantifind's Graphyte platform to support reviews of U.S. communications infrastructure risk, including ownership structures, hidden relationships, and sanctions exposure.
ThreatLocker raised $190 million in Series F funding led by Elephant to expand its Zero Trust platform, AI security controls, and international operations.
Leidos has introduced Parcata, an AI platform that uses multiple large language models to find and patch software vulnerabilities, including zero day issues.
Dymium has introduced GhostAI, a secure AI gateway that governs enterprise AI use across models, context, tools, and data through one policy engine.
Qualys has added TotalAI features for shadow AI discovery, model testing, AI agent monitoring, MCP server checks, and governance evidence. TotalAI is generally available.
Cyware and Armis from ServiceNow have partnered to connect Armis asset visibility with Cyware threat intelligence contextualization and workflow automation.
PortSwigger has launched Burp AT in public beta for Burp Suite Professional users. The tool lets pentesters assign investigative tasks to AI agents while Burp Suite controls scope, permissions, approvals, and evidence records.
Weekly newsletter about AI in Cybersecurity.