Amazon Becomes The Newest Member Of Nvidia-Led Open Secure AI Amid Push For Agentic Cybersecurity Guidelines

With Amazon joining the cohort, Nvidia’s AI security alliance has now expanded to more than 120 members after launching with just 40 names initially.

• The alliance and the Linux Foundation introduced SAFE, a framework for confidentially sharing agentic AI security incidents and near-misses across the ecosystem.
• SAFE proposes collecting and analyzing incidents, notifying affected organizations, identifying repeated control failures, and publishing evidence-based recommendations.
• Additional members, including Okta, Palo Alto Networks, Cloudflare, Microsoft, Visa, and Uber, each contributed open-source tools addressing different parts of AI agent security.

Nvidia (NVDA)‑led Open Secure AI Alliance added Amazon (AMZN) to its now 120‑plus member roster on Tuesday, as the group announced that it was working with the Linux Foundation to create a new framework for addressing cybersecurity risks tied to agentic AI.

Dubbed the Shared AI Findings Exchange (SAFE), the alliance said that the framework was aimed at turning agentic AI cybersecurity incidents and “near misses” into shared protections across the ecosystem.

The guidelines propose confidentially collecting and analyzing AI security incidents, notifying affected organizations, identifying repeated control failures, and publishing evidence-based recommendations to reduce broader risks.

“Cybersecurity is a race without a finish line,” the alliance said. “Every major technology shift has created new potential attack surfaces… When trusted ecosystems share threat intelligence openly, collective defense becomes a force multiplier.”

Nvidia’s Role In SAFE

Alongside SAFE, the alliance outlined open-source security tools several members have contributed. Nvidia's own contribution, the Nvidia Labs Object-Oriented Agent (NOOA) research harness, is designed to make agent behavior easier to test, trace, audit, and govern. 

NVDA stock rose 1.7% in morning trade amid a broader rally in the Nasdaq 100. The ETF that tracks the tech-heavy index, the Invesco QQQ Trust Series 1 (QQQ), was up over 2% in the morning hours. On Stocktwits, retail sentiment around the AI bellwether remained in ‘bullish’ territory over the past day.

CrowdStrike (CRWD) is fine-tuning Nvidia's Nemotron Nano model for cybersecurity use, reporting 96% accuracy in generating investigation queries within its Falcon LogScale platform. Meanwhile, Cisco (CSCO) contributed DefenseClaw, an agentic governance layer built on Nvidia's OpenShell, along with security-focused small language models and its Project CodeGuard initiative.

Amazon’s Entry Into Open Secure AI Alliance

Amazon, the newest joinee, contributed Strands Agents, an open-source toolkit for building AI agents, along with Cedar, an open-source authorization language designed to define and verify what actions AI agents are permitted to take.

AMZN stock bucked the broader market rally, falling nearly 2% in morning trade. Retail sentiment around the technology giant on Stocktwits remained in ‘extremely bullish’ territory over the past day.  

The drop in Amazon shares comes after Bloomberg Law reported that New ‌Jersey Attorney General Jennifer Davenport sued Amazon. The antitrust allegations stated that the online ​retailer uses ⁠its dominance to impose low pay and poor ​conditions on ​drivers. 

The Trump administration is also reportedly drafting a ban on U.S. imports of new models of Chinese data center components, which could make things more expensive for American hyperscalers, like Amazon.

How The Rest Of The Alliance Is Pitching In

Okta (OKTA) is building reference implementations for agent identity management using its Cross App Access protocol, while Palo Alto Networks (PANW) has open-sourced Agent Guard and Agent Watch from its Idira platform. 

Cloudflare (NET) contributed a Vulnerability Discovery Harness for AI agent systems, while Microsoft's (MSFT) AI Red Team open-sourced PyRIT, a Python toolkit for automated red-team testing. 

Visa (V) added its Vulnerability Agentic Harness, and Uber (UBER) open-sourced parts of its Agentic AI Detection and Response system, which the company says already tracks activity across more than 200,000 agent sessions daily on 30,000 endpoints.

Nvidia launched the Open Secure AI Alliance last month, making the debut alongside CEO Jensen Huang’s first-ever post on X. At the time, the alliance has around 40 members, including Microsoft, IBM, SpaceX and others. Within a week, that number has grown three-fold.