Investors in CrowdStrike all recoiled at one point in late March. After Anthropic revealed its Claude Mythos model, a Reddit thread titled “Cybersecurity stocks plunge as Anthropic’s ‘Claude Mythos’ leak sparks AI fear” received nearly 300 upvotes overnight. Forty-eight hours later, the stock was being dumped as though the company’s purpose had been subtly revoked. At first glance, the reasoning seemed reasonable enough: if AI can identify vulnerabilities on its own, why pay CrowdStrike to do it? However, that’s the kind of reasoning that usually appears clear for a week or so before becoming embarrassing forever.
It didn’t sit well with Jim Cramer. He presented the claim in his typical unvarnished manner on the April 21 episode of Mad Money: AI isn’t devouring cybersecurity stocks, but rather boosting them. The logic is pragmatic, almost unremarkable. With each new AI deployment, there are more surfaces to protect, more model endpoints to harden, and more potentially exploitable agentic workflows. Adding more software to an enterprise does not reduce the need for security. You enlarge it. Since February, Cramer has been making this argument in various ways, and for six weeks, the market pretended not to pay attention.
Watching this play out, it’s amazing how well the fundamentals support him. With net new annual recurring revenue of $330.7 million, CrowdStrike’s fourth-quarter revenue was $1.305 billion, up 23%. This is a 47% increase that doesn’t typically occur at businesses whose clients are quietly defecting. The Falcon Flex ARR increased by more than 120% annually. Currently, half of CrowdStrike’s users run six or more modules, and almost 25% run eight. That doesn’t fit the description of a company losing control. It’s a platform’s profile that delves deeper.
The Anthropic narrative proved to be nearly the opposite of what investors had anticipated. Rather than eliminating CrowdStrike, Anthropic announced a collaboration with it via Project Glass Wing, which is intended to safeguard both Palo Alto Networks and Anthropic’s own users. Cramer’s assertion that Anthropic requires CrowdStrike rather than the other way around has held up over time. George Kurtz, the CEO, even asked Claude directly if it could develop a tool to take the place of CrowdStrike. Kurtz is skilled at making fun of detractors without raising his voice. The model deteriorated. Investors will likely remember the small comedy in that exchange for a long time.
Since then, the movement of the stock has resembled a gradual correction of a misreading. In the week ending April 21, CRWD increased by about 13%, from about $398 to $449. It is currently close to $466, up 26% from the previous year. The performance of those who bought into the Anthropic panic and waited it out suffered significant harm. With a consensus target of $489.86, there are 42 buy ratings compared to 14 holds and not a single sell, suggesting more upside from here.
It’s difficult to ignore how frequently these panics about AI disruption follow the same pattern. The narrative surrounding an incumbent collapses for a few sessions after the announcement of a new model, but when earnings are revealed, the story subtly resets. Despite all the noise, cybersecurity is a more difficult industry than most. The dangers increase. The connections become more profound. Customers congregate on a few platforms. Although it’s still unclear how markets will react to the next AI release, Cramer appears to have read the room and the math better than most in this particular round.
