As AI agents break free, tech giants are racing to turn cybersecurity into a core feature of their businesses
Major software and infrastructure providers are embedding security features directly into their enterprise platforms to protect from rogue AI-agent threats.
A wave of artificial-intelligence adoption has come with an undesirable side effect: An explosion of cyberthreats.
AI agents can churn out code with security flaws or be weaponized by bad actors to hack into computer systems. They might even act out on their own - as was the case earlier this month, when OpenAI reported that one of its AI agents escaped its internal testing environment and carried out a cyberattack against the AI platform Hugging Face.
On Thursday, Anthropic reported a similar mishap whereby a setup error allowed its Claude models to reach the open internet during safety evaluations and compromise the infrastructure of three real-world organizations.
"The ability to write novel malware that's hard to detect has gone up," Jared Sleeper, a longtime software investor and partner at Avenir Growth, told MarketWatch. "It's easy to imagine that in six to 12 months, every application that's facing the internet in some way is being probed by a genius-level cybersecurity attacker in the form of an agent."
Tech companies are noticing, and they're increasingly making cybersecurity a core feature of their platforms. It's a trend that's unfolding across Big Tech players, software providers and financial-services entities.
Cybersecurity was a key focus on recent earnings calls. Hyperscalers showed off AI models specialized for cybersecurity, with Alphabet $(GOOGL)$ $(GOOG)$ touting its lightweight Gemini 3.5 Cyber model and Microsoft $(MSFT)$ announcing its first in-house cybersecurity model, MAI-Cyber-1-Flash. Both of these models outperformed Anthropic's Mythos on the CyberGym cybersecurity evaluation - a sign that cybersecurity capabilities are being packaged into specialized, smaller models.
Google and Microsoft have both made major cybersecurity investments, Bernstein analyst Peter Weed noted. Microsoft has steadily scooped up specialized, midsized cyber companies over the past decade, while Google made the largest acquisition in its history by purchasing AI cloud and security platform Wiz earlier this year.
"I believe Microsoft maybe has as much revenue in cybersecurity as the entire cybersecurity industry put together outside of it," Weed said. Microsoft hasn't publicly disclosed its cybersecurity revenue since 2023, when the business achieved $20 billion in annual revenue. In comparison, major pure-play cybersecurity firms Palo Alto Networks (PANW) and CrowdStrike (CRWD) reported $9.22 billion and $3.95 billion in revenue, respectively, for fiscal 2025.
This week, Microsoft also introduced Project Perception, a cybersecurity system through which agents investigate cyberthreats and fix vulnerabilities. Patrick Walravens, head of technology equity research at Citizens JMP, explained that this is another sign of Microsoft's heightened interest in security.
Cybersecurity stocks have been a bright spot amid a relentless software selloff this year. Shares of the Global X Cybersecurity ETF BUG hit an all-time high in early July, and CrowdStrike's stock has risen 68% so far in 2026. Meanwhile, the iShares Expanded Tech-Software Sector ETF IGV is down 8% since the year began.
But cybersecurity is no longer a standalone product. As AI agents become more widespread, "enterprises must approach AI software supply-chain risk in a way that is distinct from traditional application security," according to Forrester's "Top Cybersecurity Threats in 2026" report. Effective cybersecurity solutions will require integration across security, IT, architecture, data science and business governance.
As a result, enterprise software companies are increasingly doubling down on cybersecurity, according to Bernstein's Weed. "The instant that you're a platform company," cybersecurity "becomes something that your customers expect of you," he said.
Perhaps one of the biggest examples of this trend is ServiceNow (NOW), where CEO Bill McDermott has been spearheading a cybersecurity transformation of the workflow-management platform.
"In the next few years, cybersecurity for ServiceNow may be as big as ServiceNow is today," McDermott told MarketWatch.
The company has also utilized mergers and acquisitions to bolster its cybersecurity offerings, closing acquisitions for Veza and Armis earlier this year. Armis, purchased for $7.75 billion, was ServiceNow's largest acquisition in its history.
The type of cybersecurity services ServiceNow offers is materially different from the endpoint security that CrowdStrike specializes in or the network-security solutions that Palo Alto Networks sells. ServiceNow is focusing on security at the management layer attached to the platform, allowing companies to streamline threat detection and increase visibility across all AI applications. Users can also create and orchestrate response workflows, automating the process and reducing response times.
ServiceNow's AI Control Tower solution is accelerating business for the company, Citizens JMP's Walravens told MarketWatch. "That solution includes important security and risk-management functionality, which is top of mind for executives given the capabilities of new AI models like Mythos," he noted.
Legacy software company IBM is also exhibiting cybersecurity momentum. While IBM shares $(IBM)$ experienced a tough selloff in July after the company lowered revenue guidance ahead of its earnings report, cybersecurity services proved to be a notable revenue growth driver in the second quarter.
Mizuho analyst Dan Dolev told MarketWatch that cybersecurity is becoming increasingly lucrative for Mastercard $(MA)$ due to developments in agentic AI as well as AI commerce, which have also spurred upticks in malicious transactions. The company's value-added services unit, which is home to Mastercard's fraud detection, identity and threat-intelligence tools, grew 20% in the second quarter relative to a year before.
"I think a lot of people were dismissive of the growth trajectory for Mastercard, and cybersecurity has opened up a whole new avenue of growth," Dolev said. "The more risk there is of fraud, the more Mastercard gets used, and they get paid for that."
-Christine Ji -Hannah Pedone
Comments