Agent Security Surges as Obsidian Hits $1.1 Billion

In partnership with

Want to get the most out of ChatGPT?

ChatGPT is a superpower if you know how to use it correctly.

Discover how HubSpot's guide to AI can elevate both your productivity and creativity to get more things done.

Learn to automate tasks, enhance decision-making, and foster innovation with the power of AI.

Download the free guide

Agent Security Produces a New $1.1 Billion Unicorn

Obsidian Security raised an $85 million Series D at a $1.1 billion valuation, led by Crescent Cove with Greylock and Menlo Ventures participating. Nearly 70% of Obsidian customers already allow agents to interact with business data, according to the company. The funding shows that monitoring agent permissions, behavior and data access is becoming a major standalone security category.

AI Vendors Propose a Shared Incident Exchange

The Open Secure AI Alliance, which includes more than 120 organizations, opened a request for comments on the Shared AI Findings Exchange, or SAFE. The proposed framework would confidentially collect agent incidents and near misses, warn affected organizations and identify recurring control failures. Participants include NVIDIA, Cisco, CrowdStrike, Hugging Face and Red Hat, suggesting that agentic AI may gain a threat-sharing system similar to traditional cybersecurity intelligence networks.

Agentic Commerce Gets a Trust and Standards Forum

The Secure Technology Alliance launched the Agentic Trust and Commerce Forum to address identity, authentication, consent, accountability and interoperability for transactions initiated by AI agents. The group cites projections that U.S. agentic commerce could exceed $300 billion by 2030. The initiative reflects a growing realization that payment rails alone will not be enough—merchants and consumers also need standards proving what an agent was authorized to do.

Stellar Cyber Reports 99.7% Agreement With Analysts

Stellar Cyber released customer-trial results covering 138,475 alerts over 124 days, with its auto-triage agent reaching the same verdict as human analysts 99.7% of the time. The agent closed 64% of false positives and escalated 15% of alerts as true positives, reportedly returning about 19 analyst minutes per hour. The results provide unusually concrete production evidence for supervised autonomous security operations.

Fixify Finds Agents Execute One-Third of IT Actions

Fixify analyzed 17,929 agent plans, 147,351 planned actions and 52,689 skill executions across more than 40 companies. Agents performed roughly one in three IT actions under analyst supervision, while the share of actions requiring human intervention fell from 27% to 16%. The report also found that approximately three-quarters of automation failures stemmed from data problems, showing that clean identity and systems data may matter more than additional model intelligence.

Norstella Turns Biopharma Data Into Finished Agent Work

Norstella launched Atlas, an agentic platform that converts its proprietary clinical, commercial, real-world and market-access data into competitive analyses, feasibility assessments and launch strategies. Its first agent, Atlas CI, can assemble cited competitive landscapes and company profiles in minutes rather than days or weeks. Norstella’s advantage is not simply its model—it is the trillions of linked industry data points supplying the agent’s context.

Master Claude AI (Free Guide)

The professionals pulling ahead aren't working more. They're using Claude.

Our free guide will show you how to:

• Configure Claude to be the perfect assistant

• Master AI-powered content creation

• Transform complex data into actionable strategies

• Harness Claude’s full potential

Transform your workflow with AI and stay ahead of the curve with this comprehensive guide to using Claude at work.

Get The Free Guide

🗺️ The Map of AI Agents (Live & Growing)

We’re charting the entire AI agent ecosystem — thousands of options across categories.
Your next agent is already on the map.

👉 Explore the Landscape Map →