Mythos rejuvenated cybersecurity. Earnings put the rally to the test

Anthropic’s Mythos model has offered cybersecurity companies a much-needed lifeline in the age of artificial intelligence.

But this week’s cybersecurity earnings were a harsh reminder that sometimes good isn’t good enough, even with tailwinds. CrowdStrike And Palo Alto Networks lost 8% and 3% respectively.

“People are probably a little out of their depth,” said software analyst Joseph Gallo of Jefferies. “They both gave accelerator guidelines, but at the end of the day… a lot of these AI benefits take time, and it’s a multi-year process.”

Cybersecurity stocks sold off earlier in the year due to concerns that new artificial intelligence tools that can create applications at lightning speed would disrupt business models, and thus every software company.

The release of Mythos, a model deemed too powerful to release because the software’s vulnerabilities could be easily exploited, renewed interest in the industry and caused shares of CrowdStrike and Palo Alto Networks to rise more than 70% between April and the end of May.

Both companies were early partners in Anthropic’s proprietary Project Glasswing testing program; The AI ​​lab expanded to 150 additional partners this week. rubric And defensible.

This quarter’s earnings were the first major test for the Mythos-led rally, and both cyber giants’ upbeat results and aggressively optimistic AI comments were not enough for investors demanding immediate signs that AI would improve unexpectedly.

The concept that good is not enough is not a new phenomenon on Wall Street. Even AI my darling Nvidia succumbed to this phenomenon after failing to meet high estimates.

Investors have turned to earnings with a similar sentiment and the hope that AI tailwinds will help cyber companies take their earnings out of the ordinary.

Wall Street finds much to be happy about in these pressures, Gallo said, but investors may be overlooking the idea that it could take months for these headwinds to pay off.

Typical enterprise sales cycles last nine to 12 months; This means that most signs of the rise in artificial intelligence probably won’t appear until calendar year 2027. Gallo added that the fourth quarter of the calendar year is typically the strongest buying season for customers as businesses reset their budgets for the new year.

If an organization “has launched an AI product in the last quarter or two, I don’t think it’s fair to expect such a big increase now.”

The CEOs of the world’s largest cyber companies have made this point clear.

Palo Alto CEO Nikesh Arora told analysts this week that demand has been unusual during the Mythos era, with more than 1,200 companies reaching out to the cybersecurity firm to discuss AI strategy. The company has so far held 800 meetings in the last six weeks, and Arora said he has conducted nearly 100 of them.

While demand patterns are showing positive signs, analysts shouldn’t expect a sudden “windfall” next quarter as businesses buy into cyber, but expect “strong growth,” he said.

“I’m not going to get in the way and start throwing tricks and the kitchen sink at cybersecurity companies because there’s still a process, there’s a mechanism, there’s a cycle that people have adopted, and there’s implementation and deployment,” he said.

CrowdStrike CEO George Kurtz echoed a similar sentiment.

The company increased fiscal 2027 net new annual recurring revenue growth due to the negative impacts of artificial intelligence.

AI detection and response, or AIDR, is a huge new segment that could eclipse the endpoint security market but is only in its “early start” stage, Kurtz told analysts in an earnings call. He added that the company’s second-quarter pipeline already exceeds $50 million.

“Once it really becomes widespread and all companies adopt this to all their employees and workloads, I think you will see a new surge in incremental opportunities,” he said.