Artificial Intelligence / AI
'Patchpocalypse': Record Software Flaws Found by AI Send CrowdStrike, Palo Alto Stocks Soaring Since April BoliviaInteligente / Unsplash

Artificial intelligence is finding software bugs faster than companies can fix them. That is driving a surge in spending on cybersecurity and pushing shares of security companies to big gains this year.

The number of newly reported software security flaws in the U.S. National Vulnerability Database hit a record of nearly 17,500 in October, according to Yahoo Finance calculations. That is about four times the level of a year earlier.

More than 100,000 flaws have been reported so far in 2026, already more than in 2024 and 2025 combined.

A flood of fixes

Ben Emons of research firm FedWatch Advisors has given the trend a name: the "Patchpocalypse."

"A flood of AI-found bugs means a flood of patches, and every patch is a chance for something to break," Emons wrote Wednesday.

The warning comes after weeks of high-profile outages at major companies, including Microsoft, JPMorgan Chase and Apple. Each was blamed on internal problems rather than hackers.

AI models uncover old bugs

According to Yahoo Finance, the surge in discovered flaws took off after April 7, when Anthropic unveiled Claude Mythos Preview, an AI model so effective at finding vulnerabilities that the company limited its release to a small group of partners.

Other tech giants are reporting similar results. Mustafa Suleyman, who runs Microsoft's AI unit, said the company's models are finding flaws "which humans haven't found for years or even decades."

Each discovered flaw must be patched, which forces companies to spend more on security tools and staff.

Cybersecurity stocks soar

Investors have rewarded the companies that help find and fix vulnerabilities. Since April 7:

CrowdStrike: up about 124%. Palo Alto Networks: up about 100%. First Trust Nasdaq Cybersecurity ETF: up nearly 50%, more than twice the gain of the broader technology sector.

The cybersecurity fund has also held up better during the recent tech sell-off. Since the Oct. 6 peak for tech stocks, it has fallen about 11%, compared with nearly 20% for the tech sector.

Palo Alto points to patching risks

Palo Alto Networks rose sharply Thursday after raising its full-year outlook, lifting other cybersecurity stocks. CEO Nikesh Arora told analysts that "unintended consequences of patching and configuration changes" were partly behind recent outages.

Echoes of Y2K

The trend has drawn comparisons with Y2K, the late-1990s scramble to fix computer systems that stored years as two digits. A 1999 Commerce Department estimate put the cost of that effort for U.S. companies and governments at more than $100 billion.

There is a key difference: Y2K spending ended when the year 2000 arrived. The current bug hunt has no deadline, because AI models keep getting better at finding flaws.

Security operations in focus

Investors see security operations centers, the teams that monitor networks around the clock for attacks, as a growing area of spending. Somesh Dash, a general partner at venture firm IVP, which is an Anthropic investor, sees this year's race in that area.

Anthropic debut

The trend comes as Anthropic makes its Nasdaq debut after raising $81.3 billion in the largest IPO on record. While cybersecurity companies have already benefited from the AI bug boom, Anthropic's new shareholders are still waiting to see how its stock performs.

Disclosure: This article was produced with the help of Claude, an AI model developed by Anthropic, a company mentioned in this article.

This article is for informational purposes only and does not constitute financial advice.