By Britney Nguyen

The software sector outperformed the chip sector to a historic degree in the face of escalating AI fears

CrowdStrike's stock reached an all-time closing high Monday.

Cybersecurity stocks were the top performers in the S&P 500 on Monday, showcasing how the technology trade has become starkly divided in the face of escalating warnings about artificial intelligence.

Over the weekend, Anthropic CEO Dario Amodei outlined his concerns about the accelerating speed of AI development - warning about the potential of models soon reaching a point where they are so powerful that they could spur legions of bots to take over the entire internet. He called for major players to pace themselves and focus on safer AI progress.

That commentary weighed on chip stocks Monday, but it was a boon for the cybersecurity sector. Shares of CrowdStrike Holdings (CRWD) rose 13.8% Monday, marking an all-time closing high, according to Dow Jones Market Data. Palo Alto Networks' stock (PANW) gained 13% for its largest move since April 2025, while Fortinet's stock (FTNT) was up 9% at the close.

Cybersecurity stocks helped the broader software sector put up a historic performance relative to the chip sector. The iShares Expanded Tech-Software Sector ETF IGV trounced the iShares Semiconductor ETF SOXX by 10.67 percentage points Monday, which Dow Jones Market Data noted was the largest single-day outperformance in history.

In a blog post, Anthropic's Amodei cited his worries about recursive self-improvement, or RSI, which involves an AI model reaching the point of being able to train and improve itself. He also wrote about the recent attack on open-source-model repository Hugging Face by AI agents built by OpenAI. Amodei received support for his stance from OpenAI CEO Sam Altman as well as Tesla (TSLA) and SpaceX (SPCX) CEO Elon Musk.

Regardless of how fast AI agents roll out to enterprises, "those agents will still need to be secured, governed and observed," and will therefore maintain "the structural demand backdrop for cybersecurity and select infrastructure names," even if AI training slows down, Evercore ISI analyst Kirk Materne said.

In his view, "greater scrutiny around AI safety could increase enterprise attention on identity, data governance, observability and security," Materne said in a Monday note to clients. He noted that AI-related benefits for most infrastructure software companies will come from user adoption and the rise of inference, or the process of AI models making predictions based on what they've learned.

For cybersecurity companies specifically, Materne said the fearful commentary on AI "further reinforces the durability of security spend," given cybersecurity's growing role in enabling AI deployments. Therefore, he sees Okta (OKTA), SailPoint (SAIL), Palo Alto Networks and CrowdStrike benefiting as companies prioritize identity security software.

Materne also pointed out that cybersecurity vendors are expanding partnerships with AI labs, particularly for run-time security, which involves protecting actively running systems.

Across cybersecurity names, Materne said Palo Alto Networks and CrowdStrike "remain best positioned to capture" growing spending on cybersecurity over time, given the companies' "ability to deliver integrated visibility and response across the broader security stack."

Don't miss: Anthropic's safety-first image takes a hit ahead of its massive IPO: 'AI could kill all humans'

Jefferies analyst Joseph Gallo said that although momentum for agentic AI security is building, it's still early. He expects "initial signals to emerge" later this year, while "material contribution" will likely come in 2027 and beyond.

Like Materne, Gallo said in a Sunday note that he expects identity-security vendors to be among the first beneficiaries of inflecting agentic security efforts given "visibly healthy demand" and growing adoption of AI agents.

Although he agreed that cybersecurity risks "are real," D.A. Davidson managing director Gil Luria said companies should address that by "hardening code bases," or strengthening software to handle attacks, rather than by slowing down AI development.

Apart from cybersecurity companies, Luria said in a note to clients that Microsoft (MSFT), Amazon.com (AMZN) and Google (GOOGL) (GOOG) "would be great candidates to lead" that effort, given their large customer bases.

-Britney Nguyen

This content was created by MarketWatch, which is operated by Dow Jones & Co. MarketWatch is published independently from Dow Jones Newswires and The Wall Street Journal.

(END) Dow Jones Newswires

09-14-26 1711ET