The resurgence of cybersecurity groups can only be described as surprising. In just a few months, CrowdStrike and Palo Alto Networks have gone from being obsessives in the hated software industry to full-fledged beneficiaries of artificial intelligence. From the beginning of the stock market crash earlier this year, Jim Cramer has been shouting from the rooftops that cyber shouldn’t be lumped in with run-of-the-mill enterprise software. That’s why we bought so much of CrowdStrike in March, just after the club’s stock hit a 52-week low of about $86 per share in late February. While we supported this issue on CrowdStrike, we downplayed Palo Alto Networks, which made some adjustments in April and is now about $40 a share above its 52-week low of $139 in late February. Fast forward to this week, CrowdStrike and Palo Alto Networks hit intraday highs on Monday before falling back a bit on Tuesday. While it’s good to see stocks recovering and reaching new heights, we need to be mindful of the speed of these movements. While we may think CrowdStrike is a great company and the share price will rise further from here, that doesn’t mean we’ll leave it alone until it has a significant impact on the risk/reward profile of the club’s overall portfolio. On Monday, we returned our weight in CrowdStrike to around 4% and posted a healthy profit. Last week, we devalued Palo Alto Networks stock, but recent gains have brought the stock’s weight back to around 2%. CrowdStrike is up 68% since the beginning of the year, while Palo Alto Networks is up 88%. So how did the market finally understand what we’ve known all along? At a high level, this move may not be all that different from what has played out in memory chip stocks over the past few months. If you think about it, it was precisely the software representation of memory trading that drove much of this year’s market gains. Both are tied to AI and the insatiable need to adopt and implement it. On the hardware side, memory stocks are extremely hot. SanDisk is up more than 625% since the beginning of the year, while Micron Technology is up 250%. Recognize: The world may need more central processing units (CPUs) to run agent AI workloads at scale than one might think, but the real bottleneck in building infrastructure turns out to be the availability of high-bandwidth memory. However, even after the hardware is installed and the software applications are ready for deployment, organizations need to ensure that they can implement AI securely without compromising not only their customers’ data, but also their own data. In today’s world, data is the new gold. This means that data breaches are the modern equivalent of bank robberies. However, the bad guys don’t just get away with one hit. They can try, fail, and keep trying. In fact, AI allows cybercriminals to attack a million times a day, making it much easier for cybercriminals to evade capture when they are behind a computer screen, often half a world away from their targets. Trying to implement powerful new technologies such as artificial intelligence is no small challenge. This is especially true for agent AI systems that give permissions to thousands of digital agents needed to complete tasks autonomously. All such access exponentially increases vulnerabilities and puts them in the hands of hackers. It’s an arms race in which AI will be used for both defense and offense. Indeed, businesses realize they have no choice but to deploy AI. The failure to adapt to the AI era is similar to the lack of mobile apps after the launch of the iPhone, or the failure of brick-and-mortar businesses to establish an online presence as it becomes increasingly clear that Amazon is ushering in permanent changes in the way consumers shop online. Either evolve and adapt, or be destroyed. This new technology is so powerful that it’s up to cyber companies to decide how quickly it can be deployed. CrowdStrike CEO George Kurtz made a similar point during an earnings call in June, saying, “The inflection point is that all the players in this value chain are experiencing hypergrowth, and all of these technologies require cybersecurity.” Kurz added, “The mythical moment crystallized a market reality, and for the first time in my career, the market’s view of the role of cybersecurity has gone from being viewed primarily through the lens of risk management, compliance, and protection to being recognized as a strategic accelerator and key enabler of AI adoption.” That “Mythos moment” happened in early April when Anthropic announced Project Glasswing. The project initially brought together 11 organizations, including CrowdStrike and Palo Alto Networks, to secure the startup’s new Mythos model. This model has been tested and proven to be good at finding security holes. It was at this time that cyber stocks began to take off in earnest. Since then, Project Glasswing has expanded to approximately 200 organizations. With that in mind, there is one important difference between cyber trading and memory trading. Memory chips represent hardware sales, and sales can be somewhat volatile and cyclical as supply and demand dynamics play a large role in pricing, despite efforts by companies like Micron to reduce that cyclicality with long-term contracts. Cybersecurity, on the other hand, is an ongoing and recurring revenue stream, the kind that Wall Street offers big rewards. It is also important to distinguish because cybersecurity deals are about volume growth over time, whereas hardware deals are about physical bottlenecks that lead to significant pricing power for the companies selling the hardware. Like many hardware players, you shouldn’t expect a surprising windfall in revenue for cybersecurity companies. For example, Micron’s earnings per share last quarter rose an otherworldly 1,200% year over year. Palo Alto’s EPS increased by 156% and CrowdStrike’s increased by 51%. The huge increase in hardware revenue is the result of buyers competing for a limited supply of physical goods (remember, more money chasing the same or less supply is how you create inflation). In software, these dynamics don’t really exist. Software allows companies to scale with little increase in marginal cost. A customer is looking to implement a security solution on the hardware they already have in place. The more hardware is installed and the more workloads run on the hardware over time, the greater the opportunity for cybersecurity companies. This is also why we think these companies are names to own for the long term as AI becomes more pervasive. However, the newly installed hardware on which AI applications will eventually run is intended to expand the addressable market rather than to increase short-term revenue. Palo Alto Networks CEO Nikesh Aurora made a similar point during the company’s latest earnings call, saying, “If you think, like many SaaS companies, that the terminal value of cybersecurity is gone, this terminal value is here to stay. In reality, we’ve just put a long-term ‘G’ in our long-term growth rate model for cybersecurity.” He added: “I’m not going to get ahead of myself and blame the numbers for cybersecurity companies, because there are still processes and mechanisms and cycles that people buy into, and execution and deployment. So, to that extent, do I see good demand? Yes. To what extent do I see this demand lasting much longer? Yes. To what extent can we expect to wind up next quarter, the next quarter? No, I expect solid growth,” he added. Therefore, investors in cyber companies do not need to worry too much about investment cycles. Hyperscalers Amazon, Alphabet, Microsoft, and Metaplatforms are expected to deploy a combined $750 billion this year, up more than 80% from 2025. If anything, the real demand for cyber companies will come after the AI infrastructure is built and workloads are running on the installed hardware. In that world, AI will become more pervasive, capabilities will become more cutting-edge, and attacks from malicious actors and the associated demands on cybersecurity will continue unrelenting. (Jim Cramer’s Charitable Trust is long CRWD, PANW, AMZN, GOOGL, MSFT, META. See here for a complete list of stocks.) As a subscriber to Jim Cramer’s CNBC Investment Club, you will receive trade alerts before Jim makes a trade. After Jim sends a trade alert, he waits 45 minutes before buying or selling stocks in his charitable trust’s portfolio. If Jim talks about a stock on CNBC TV, he will issue a trade alert and then wait 72 hours before executing the trade. The above investment club information is subject to our Terms of Use and Privacy Policy, along with our disclaimer. No fiduciary duties or obligations exist or arise from your receipt of information provided in connection with the Investment Club. No specific results or benefits are guaranteed.
