The AI Boom Has a Security Problem – and 2 Potential Winners

Listen to the audio version of this article (generated by AI).

Editor’s Note: As AI agents like Meta’s newly released Muse take on more work, companies face a growing challenge – making sure those agents follow the rules. A recent review of AI agents running on Chinese AI models found they did something similar to what recently happened with U.S. models like ChatGPT. Researchers found the agents deceived and tested boundaries.

For investors, the questions raised are practical: Which businesses could get paid to make these systems safer?

My friend and colleague Jonathan Rose has two stocks in mind (which he’ll show you below.) Jonathan brings nearly three decades of trading experience on some of America’s busiest floors.

For Jonathan, finding a tech company positioned to profit is only the start. He then studies factors like unusual trading activity and weighs the risks before making a trade. His Advanced Notice recommendations have averaged an 81% gain since inception in 2024 – that’s winners and losers.

Today, Jonathan explains the budding opportunity he’s seeing – and what he’s watching before putting money to work. You can follow his analysis on Masters in Trading LIVE, free on YouTube every market day at 11 a.m. Eastern. Sign up here for the daily link.

So let me turn things over to Jonathan…

********************

Hey, guys.

Just after midnight on January 25, 2003, a tiny piece of code began racing across the internet.

It was a computer worm – dubbed the SQL Slammer— that could copy itself from one vulnerable machine to the next without anyone clicking a link.

Slammer was only 376 bytes long. It simply told each infected server to send out more copies of itself, as fast as it could.

That was enough.

The number of infected machines doubled about every 8.5 seconds. Within 10 minutes, Slammer had reached more than 90% of the computers vulnerable to it. The flood of traffic knocked networks offline, interfered with ATMs, and disrupted airline flights.

Microsoft Corp. (MSFT) had released a fix for the weakness Slammer exploited months earlier, but plenty of organizations still hadn’t installed it.

We’re at a similar moment with AI agents—software that can search for information, use tools, and carry out tasks with little human guidance. Companies putting agents to work have to control what those agents can access and catch them when they do something nobody intended.

Recent incidents have shown why. OpenAI says its reviews have found agents from their lab bypassing access controls, using exposed credentials, and interacting with third-party systems beyond their intended access. It has notified dozens of affected hacked parties – including the Australian and U.S. governments – while the reviews continue.

That creates a new job for cybersecurity companies. And I see two stocks investors should consider buying now.

Let me show you what each company sells—and what I’m watching before making a trade.

AI Agents Need Someone Watching the Door

Think of an employee’s credentials as a set of keys. An AI agent working for that employee may need access to certain files and applications. Give it too many keys, and an error—or an attacker manipulating the agent—can carry it somewhere it shouldn’t go.

That makes identity security unusually important.

This is where Palo Alto Networks Inc. (PANW) has made a significant move. It completed its acquisition of CyberArk in February, adding that company’s identity-security products to a business that already sells network, cloud, and security-operations tools.

In plain English, Palo Alto is trying to give large customers a single place to manage more of their defenses, including the permissions granted to AI agents. The integration is still underway, so investors need to watch how well it delivers on that plan.

In its latest reported quarter, Palo Alto’s revenue rose 34% from a year earlier, to $3.41 billion. Its next-generation security annual recurring revenue—the value of subscriptions it expects to collect over a year—reached $9.1 billion. Those figures include the effects of acquisitions, so I’m watching what growth looks like as the businesses are brought together.

PANW is my first buy of the two. It gives me a large, established cybersecurity business with a clear path to sell more to customers as they add AI agents.

CrowdStrike Holdings Inc. (CRWD) comes at the AI agent problem from another direction.

Its Falcon software already operates across customers’ computers and other devices—the places where a great deal of AI-agent activity starts. This month, CrowdStrike introduced Falcon Guardian, designed to help companies discover which agents are running, see what they do, control which ones have permission to run, and respond when their behavior becomes dangerous.

It’s a new product, so I would watch customer adoption before assuming it will become a major revenue source.

CrowdStrike ended its latest reported quarter with $5.84 billion in annual recurring revenue, up 25% from a year earlier.

I would buy CRWD, too, but I’d start with a smaller position. Investors already recognize how strong this company is. That can leave less room for disappointment if a new product takes longer to catch on or growth slows.

And spending on security is still growing. Gartner forecasts worldwide information-security spending of about $244 billion in 2026, with demand for products that both use AI to improve defenses and secure companies’ own AI use.

Even if the Company Is Right, the Trade Still Has to Work

I look for a good business first. Then I look at the market.

One signal I follow is unusual trading activity: a burst of trading that stands out from a stock’s normal pattern. It can tell me that someone is making a large bet.

For example, in August, we spotted a spike in trading activity in oil-services company SLB NV (SLB). I recommended a bullish trade. Over the following week, SLB shares rose about 9%. Our recommended trade gained 219%.

We saw another example in MP Materials Corp. (MP). Our recommended trade returned 534% in three days. For both examples, the maximum possible loss was the amount paid for each contract.

Those were standout winners. These trades can also expire worthless, which is why choosing the trade and controlling its size matter as much as finding the company.

So if I see that activity in a stock, I check the news, the stock’s price, and the amount I could lose. Some days I’m looking at cybersecurity. Other days, it’s energy, healthcare, or a company I hadn’t expected to discuss when the market opened.

It’s what I do every market day on Masters in Trading LIVE. I’m on YouTube at 11 a.m. Eastern, free, walking through the stories and trading activity catching my attention and taking questions as we go.

Sign up for Masters in Trading LIVE here. Once you’re onboard, we’ll send you a daily email with the link to watch, along with more pieces like this one and other research.

You’ll be hearing more soon about my free $10K to $100K Challenge event in October, where I’ll go deeper into the signals and risk rules behind my trades. Till then, I hope you’ll check out MIT LIVE and see how I work through them while the market is open.

SQL Slammer didn’t create the cybersecurity business. It made the cost of leaving connected systems unprotected impossible to miss.

AI agents are extending that same problem to a new kind of worker. Companies need to know what their agents can reach, what they’re doing, and how to stop them if they start hacking.

I think Palo Alto and CrowdStrike have a real chance to collect a meaningful share of that spending.

Remember… the creative trader wins…

Jonathan Rose

Founder, Masters in Trading

P.S. Jonathan’s edge is the way he joins a strong investment idea to an actual trading decision. He’s given you two cyber companies to examine today, along with the questions he’s asking about each. You can watch him do that work across many more stocks on Masters in Trading LIVE, free at 11 a.m. Eastern every market day. Sign up here for the daily link. And keep an eye out for his free October $10K to $100K Challenge event.

Sincerely,

Louis Navellier

Editor, Market 360