TradingKey - Recently, Anthropic CEO Dario Amodei publicly called for slowing down the development pace of frontier AI models and proposed allowing major AI companies to coordinate a slowdown in R&D speed. OpenAI CEO Sam Altman also supported taking a more cautious approach to AI development.

As AI agents rapidly enter enterprise applications, AI is no longer limited to generating text or images, but is beginning to gain the ability to access databases, invoke software, write code, and even autonomously execute tasks. This elevation of AI privileges has further heightened the importance of model security, identity authentication, data protection, and defense against cyberattacks.

Following the news, the market reacted noticeably. On September 14, discussions of an AI slowdown weighed on select semiconductor and tech stocks, while cybersecurity stocks such as Palo Alto Networks (PANW), CrowdStrike (CRWD), and Cloudflare (NET) rose significantly, with both CRWD and PANW gaining over 13% that day.

What Is AI Safety?

AI security is not a single product or a standalone industry, but rather a suite of technologies and services built around the security risks arising during the development, deployment, and use of artificial intelligence models, AI applications, and AI agents.

Unlike traditional cybersecurity, which primarily protects servers, endpoint devices, and corporate networks, AI security introduces a significant number of new risks associated with large models.

For example, when employees use AI tools such as ChatGPT or Claude, they might inadvertently upload customer data, source code, or other sensitive information; hackers could also induce AI to bypass existing security restrictions through methods like "prompt injection." As AI agents gain access to corporate databases, email accounts, and internal software, companies must also determine which data each agent can access, what actions it can execute, and how to prevent abnormal AI behavior.

Therefore, the current AI security market can be broadly divided into several main areas:

The first is AI model and application security. This primarily includes detecting risks such as model vulnerabilities, prompt attacks, sensitive data leakage, and malicious model outputs.

The second is AI agent identity and access management. As the number of internal AI agents grows within enterprises, companies need to manage agent identities much like employee accounts and restrict the data and software they can access.

The third is AI-driven cyber defense. Cybersecurity firms are beginning to leverage AI to automatically discover attacks, analyze threats, and execute responses to address attackers using AI to increase the automation of cyberattacks.

The fourth is AI data and traffic security. Enterprises must monitor data traffic generated when employees and AI agents call external models like OpenAI and Anthropic, preventing sensitive data from leaving the corporate network.

This also explains why AI security stocks are not a homogeneous group of companies: Palo Alto Networks primarily covers AI model and agent security, CrowdStrike focuses on AI-automated cyber defense, while Cloudflare leans more toward AI traffic, gateways, and agent identity management.

Palo Alto Networks (PANW): AI Security Business Achieves Revenue at Scale

PANW's direct link to AI security primarily stems from Prisma AIRS. Prisma AIRS currently covers multiple domains including AI model scanning, AI application security, AI Gateway, Agent identity management, runtime protection, and AI Agent behavior monitoring. As enterprises gradually shift from chatbots to AI Agents capable of executing tasks autonomously, PANW is extending its traditional cybersecurity platform across the entire AI application lifecycle.

More notably, Prisma AIRS has begun generating quantifiable revenue. As of the fourth quarter of fiscal 2026, Prisma AIRS achieved an annualized recurring revenue (ARR) of approximately $120 million, having been officially commercialized for about a year, with its customer base exceeding 800 and more than doubling from the previous quarter.

During the same period, PANW's quarterly revenue reached $3.41 billion, up 34% year-over-year; Next-Generation Security ARR reached $9.1 billion, up 63% year-over-year.

PANW's AI security business is primarily built on the company's existing base of large enterprise cybersecurity clients, allowing cross-selling of Prisma AIRS with cloud security, network security, and identity security products. Therefore, the future ARR growth rate and customer count of Prisma AIRS will serve as key metrics for tracking the commercialization of the company's AI security business.

Recently, Wall Street has also adjusted its valuation expectations for PANW accordingly. RBC Capital raised its price target on PANW from $434 to $475 and maintained an "Outperform" rating, citing the latest quarterly performance, the fiscal 2027 outlook, and management's confidence in long-term free cash flow margin targets.

CrowdStrike (CRWD): Focus on AI-Driven Cybersecurity Automation

CRWD's AI security layout is focused more on security operations and threat detection.

CRWD's Charlotte AI can combine endpoint, identity, cloud, and network data to conduct threat investigations and is gradually evolving toward an Agentic SOC, where multiple AI agents automatically perform analysis, investigation, and response tasks traditionally handled by security teams. The company has also established the Charlotte AI AgentWorks ecosystem, with partners including OpenAI, Anthropic, AWS, Nvidia, and Salesforce, further expanding the application of AI agents in security operations.

According to its fiscal second-quarter financial report, CRWD generated revenue of $1.47 billion, up 26% year-over-year; subscription revenue grew 27% to $1.4 billion, ARR increased 25% year-over-year to $5.84 billion, and net new ARR added during the quarter was approximately $333 million.

Based on financial data, CRWD remains primarily driven by overall cybersecurity demand rather than relying solely on its AI security business. The company also does not separately disclose revenue from AI security products, making it currently difficult to quantify the specific contribution of AI security to its total revenue.

For CRWD, investors can focus on whether enterprise AI deployments are driving the adoption rate of new modules on the Falcon platform, and whether AI security-related products can further increase average customer spend.

Institutions have also raised their expectations for CRWD. Truist raised its price target from $245 to $300 while maintaining a "Buy" rating, believing that businesses such as AI Detection and Response, Identity, Cloud, and SIEM could support higher long-term net new ARR growth.

Cloudflare (NET): Betting on the Secure Traffic Gateway for AI Agents

NET's AI Gateway sits between enterprise applications and model providers such as OpenAI, Anthropic, and DeepSeek, enabling the monitoring of AI requests and model responses. Its Guardrails feature can inspect prompts and model outputs, flagging or blocking them based on enterprise-defined security rules.

In August this year, NET launched the Identity-Aware AI Gateway, linking AI requests to the identities of specific employees or automated agents, enabling enterprises to track AI usage, set access permissions, and identify anomalous behavior.

According to its latest financial report, NET's second-quarter revenue reached $696.1 million, up 36% year-over-year; however, its GAAP operating loss widened from the same period last year to $205.7 million.

Therefore, NET's business connections to AI traffic and AI agent security are growing, but at this stage, AI Gateway's contribution to overall revenue has not been separately disclosed. Rather than focusing solely on revenue growth, investors also need to monitor its profit margins and whether AI-related businesses can generate more explicit commercial revenue.

Notably, several institutions raised their price targets for NET. Citi sharply raised its target from $265 to $400, Goldman Sachs raised its target from $266 to $389, and BTIG raised its target from $314 to $382. RBC Capital also raised its price target from $290 to $346, primarily citing the company's accelerating growth, raised guidance, and commercialization opportunities in AI agent traffic.

PANW, CRWD, NET: Detailed Comparison of Three AI Security Stocks

Overall, for investors seeking AI security stocks, PANW offers the most comprehensive coverage; CRWD is better suited for investors focused on AI-driven cyberattacks and automated defense trends; while NET leans more toward AI Agent infrastructure and AI traffic security.

Find out more