Detailed write-ups
Google attributes the first in-the-wild AI-generated zero-day (May 11)
Google’s Threat Intelligence Group disclosed that a prominent cybercrime group used an AI model to discover and weaponize a zero-day that bypasses two-factor authentication. China- and North Korea-linked actors are deploying agentic offensive tools including Strix and Hexstrike. The 90-day disclosure-to-patch model now assumes adversaries can compress that window dramatically.
Sources: SecurityWeek · CNBC · The Hacker News
OpenAI launches Daybreak with GPT-5.5 (May 11)
Daybreak pairs GPT-5.5 with Codex Security and a partner network (Cloudflare, Cisco, CrowdStrike, Oracle, Zscaler) to find, validate, and patch vulnerabilities. Three model tiers ship with progressively stricter access controls on the cyber-tuned variants. The launch is widely framed as OpenAI’s answer to Anthropic’s Project Glasswing.
Sources: The Hacker News · MacRumors · CIO Dive · SC Media · Gizmodo
Microsoft MDASH autonomously finds 16 Windows vulnerabilities (May 12)
MDASH, a multi-model agentic scanning harness, autonomously discovered and validated 16 previously unknown Windows vulnerabilities — including critical RCE flaws in the kernel TCP/IP stack and IKEv2 — all patched in this week’s Patch Tuesday. Microsoft reports 21/21 planted vulnerabilities found with zero false positives and 88.45% on the public CyberGym benchmark.
Sources: Microsoft Security Blog
Exaforce $125M Series B (May 12)
Three-year-old Exaforce closed a Series B at a ~$725M valuation. HarbourVest, Peak XV, Mayfield, Khosla, and Seligman Ventures backed the round. The thesis: AI for catching and stopping cyberattacks in real time.
Sources: TechCrunch
Akamai to acquire LayerX for ~$205M (May 14)
Akamai is buying LayerX, a browser-based AI usage control and enterprise browser company. The deal pulls Akamai into the workforce-security AI-control category alongside Island, Talon (acquired by Palo Alto), and Menlo. Watch for downstream packaging into Akamai’s edge security stack.
Sources: Akamai
White Circle raises $11M for AI control (May 13)
Seed round to help organizations monitor, secure, and control AI model actions. Early entry in the “AI runtime control” category that vendors like Prompt Security, Lakera, and now LayerX-inside-Akamai are crowding into.
Sources: SecurityWeek
Grego AI debuts with $250K AI-found exploit bounty (May 12)
New AI security startup launched out of stealth, claiming a record $250,000 bug bounty paid for a vulnerability discovered by its AI. Worth watching as a proof point that defensive AI is now finding marketable, paid exploits.
Sources: SiliconANGLE
PANW Idira: identity for the AI enterprise (May 12)
Idira combines modern PAM with agentic-AI-aware identity governance, consolidating capabilities for legacy CyberArk customers following PANW’s $25B CyberArk close in February. Positioned as the first identity platform purpose-built for an AI-agent enterprise.
Sources: Palo Alto Networks
|
