Jensen Huang told investors on Sept. 10 that cybersecurity is likely to become AI’s next major growth market, according to a Seeking Alpha report on his comments at the Goldman Sachs Communacopia + Technology Conference.
Hours later, Anthropic told Reuters it had disrupted a Russia-linked hacking campaign that used its Claude models against more than 20 Ukrainian government and defense targets.
Huang’s prediction made headlines. The more concrete evidence sits in two earnings reports that landed before he ever said a word.
CrowdStrike (CRWD) and Palo Alto Networks (PANW) are both racing to answer the exact question Huang just raised for investors, and they are answering it in almost opposite ways.
CrowdStrike is building its AI defenses itself
CrowdStrike unveiled SafeMind at its Fal.Con conference on Sept. 1, an agentic cybersecurity system built by its own Cyber Superintelligence Lab on top of open Nemotron models, according to a CrowdStrike press release.
The system pairs an offensive model, Red Tempest, with a defensive model, Blue Solano, and runs them against each other continuously within CrowdStrike’s Falcon platform.
CrowdStrike trained SafeMind on its own Falcon telemetry, threat intelligence and 15 years of incident response data, a dataset a general purpose AI lab cannot easily replicate, according to the company.
Related: Wall Street panicked over AI. Then came an 8-figure cybersecurity twist
That is a meaningfully different bet than simply reselling someone else’s frontier model with a security label on it.
CrowdStrike says SafeMind detects threats 29% more accurately and remediates them six times faster than the frontier models it benchmarked against, according to the company’s own testing.
Those figures come from internal evaluations rather than independent verification, which matters when the company is asking investors to price in a new product category based on its own scorecard.
The approach is already showing up in demand. CrowdStrike’s fiscal second quarter revenue rose 26% to $1.47 billion, with net new annual recurring revenue climbing 51% to a record $333 million, according to the company’s earnings release.
CEO George Kurtz also disclosed an eight-figure Falcon Flex deal with a frontier AI lab, evidence that AI developers themselves are becoming CrowdStrike customers rather than just its disruptive threat.
Palo Alto is buying its way into the same race
Palo Alto Networks took the opposite path, acquiring its way into agentic AI security instead of building one flagship model.
The company closed its fiscal year having integrated CyberArk and Chronosphere and just added Console, an AI native platform built for agentic enterprise workflows, according to a press release detailing the results.
The strategy is working, at least by the numbers. Palo Alto’s Next Generation Security annual recurring revenue reached $9.1 billion, up 63% year over year, and total remaining performance obligations crossed $20 billion for the first time, rising 34% to $21.2 billion, according to the company’s earnings release.
Yet Palo Alto swung to a GAAP net loss of $282 million in the quarter, compared with a profit a year earlier, largely reflecting the cost of digesting three acquisitions at once.
CEO Nikesh Arora said AI advances are “elevating cybersecurity to the top of the CIO priority list,” according to the company’s earnings release, but the profit and loss statement is telling a more complicated story than the growth headline.
PixeloneStocker / Getty Images
Investors haven’t decided which approach wins
CrowdStrike’s beat sent shares up roughly 20% on Aug. 27, the stock’s best trading day ever, according to CNBC.
Palo Alto’s beat produced the opposite reaction days later. Revenue rose 34% to $3.41 billion in its fiscal fourth quarter, yet shares fell more than 5% that day as investors focused on margin pressure instead of growth, CNBC reported.
More AI:
- Nvidia just made a move Wall Street wasn’t ready for
- Microsoft just took sides in AI policy fight
- OpenAI just disclosed something genuinely alarming
That split suggests something more specific than a blanket AI cybersecurity trade.
A model built and owned internally produces a cleaner growth story that markets can reward immediately, while stitching together several acquisitions in one year invites harder questions about integration costs, even when the underlying revenue numbers look just as strong.
The threat behind both bets isn’t hypothetical
Anthropic’s own report, covering activity from December through August, found that a group whose tradecraft matched Russia-linked Midnight Blizzard used phishing, hijacked hotel Wi-Fi networks, and compromised WhatsApp accounts to target Ukrainian officials and people tied to the country’s drone supply chain.
Anthropic said humans mostly supervised while AI carried out large portions of the operation.
That capability jump is not unique to one company’s models. A joint study by cloud security firm Wiz and AI lab Irregular found that AI agents completed sophisticated offensive security challenges for under $50 in computing costs, versus close to $100,000 for the same work done by paid human researchers, Fortune noted.
Cheaper, faster attacks are exactly what turns a forecast like Huang’s into a line item on a corporate budget, and both CrowdStrike and Palo Alto are betting that budget grows regardless of which strategy investors reward first.
The next few quarters should settle which approach compounds faster. Falcon Flex adoption will show whether CrowdStrike’s model, built and owned in house, keeps converting AI anxiety into signed contracts at the pace it did this summer.
Palo Alto’s Next Generation Security ARR growth will show whether its acquisition strategy can match that pace once the CyberArk and Chronosphere integrations mature.
Either outcome will say more about who actually wins the market than Huang’s comment on a Thursday conference stage ever could.
Related: OpenAI’s answer to rising AI hacking risks has two tiers
