Jensen Huang Says Cybersecurity Could Be AI’s Next Major Market

Nvidia CEO Jensen Huang used Goldman Sachs’ Communacopia + Technology Conference to make a case that the next large AI market may be cybersecurity—not simply more general-purpose chatbots or coding tools. His argument is straightforward: if AI can automate software development, it can also help discover weaknesses in software and accelerate remediation. 10

That claim fits a broader Nvidia strategy. Huang is positioning the company as a supplier of the compute, networking, software, models and system design needed to build what Nvidia calls AI factories, rather than as a company that only sells GPUs. 2

Why Huang sees cybersecurity as the next AI use case

Huang said cybersecurity would “likely” be AI’s next major use case. He linked the opportunity to AI-assisted programming: faster code creation and modification can increase the pace at which vulnerabilities are found, exploited and fixed. In that environment, security teams need automation to keep up with both the volume and speed of threats. 10

The commercial logic is not that AI makes cyber risk disappear. It is that AI shifts security toward continuous, machine-speed work: analyzing code, modeling attacks, prioritizing alerts and helping defenders close gaps faster.

Huang also pushed back on alarm-driven narratives around AI. Reporting from the conference quoted him saying that the industry’s heavy cybersecurity discussion coincided with product launches, adding: “What better way to create demand than to create a problem?” 6 That framing is provocative: it questions whether some security rhetoric is commercially motivated, even as Nvidia itself is investing in AI-based cyber defense.

The product proof point: CrowdStrike SafeMind

The most concrete example is Nvidia’s work with CrowdStrike. In September, the companies announced CrowdStrike SafeMind, an agentic cybersecurity system developed through CrowdStrike’s Cyber Superintelligence Lab. 19

SafeMind combines CrowdStrike’s purpose-built models and agentic harnesses with defensive models based on Nvidia’s open Nemotron models. The design uses a continuous loop in which offensive and defensive capabilities challenge and improve one another. 19

This matters because it illustrates the kind of market Huang is describing: not a generic AI assistant applied to security, but specialized models, security data, agents and infrastructure assembled into a security product.

Cisco, CrowdStrike and Palantir are among the organizations associated with Nvidia’s Open Secure AI Alliance, which focuses on open models, agent harnesses and tooling for AI-era security. 23 Nvidia has also said that CrowdStrike and Palantir use Nemotron open models for long-running enterprise agents. 29 The available reporting does not, however, establish that every alliance member uses the same Nemotron-based product architecture.

A $3–$4 trillion infrastructure thesis

Cybersecurity was only one part of Huang’s conference message. He reiterated Nvidia’s view that global AI infrastructure spending could reach $3 trillion to $4 trillion by 2030. 32

Huang described the transition as a change in the computing model itself: moving beyond retrieval-based systems toward generative AI and large, complex models that require substantial compute and interconnect capacity. Nvidia’s role, in this framing, is to co-design AI factories rather than simply ship individual processors. 2

He said Nvidia expected roughly 70% year-over-year growth even though unconstrained demand was growing by more than 100%. The gap is important: Huang’s message was that supply—not demand—is the principal constraint. Reported bottlenecks include chips, memory, advanced packaging, land and electric power. 32

The financial context behind the claim

Nvidia’s latest results provide the scale behind Huang’s infrastructure thesis. The company reported $96.2 billion in revenue for the second quarter of fiscal 2027, ended July 26, 2026—up 18% from the prior quarter and 106% from a year earlier. 52

For the third quarter of fiscal 2027, Nvidia forecast revenue of $108 billion, plus or minus 2%, while assuming no Data Center compute revenue from China in that outlook. 52

Those figures should not be confused with fiscal 2026 results. The $96.2 billion quarter was fiscal Q2 2027.

What the available reporting does not prove

Several claims often grouped into accounts of Huang’s remarks need more evidence than the supplied reporting provides. There is no substantiated primary-transcript basis here for a specific criticism by Professor Erik Gordon, a definitive statement that OpenAI views cybersecurity as a major commercial opportunity, or Huang’s detailed conference response to allegations of circular financing through customer investments.

Those questions are material, particularly as AI infrastructure financing grows. But absent a direct transcript, company filing or reliable reported quotation, they should not be treated as settled facts.

The takeaway for security and AI buyers

Huang’s main point is less about a single cybersecurity product than about market structure. As AI takes on more programming and operational work, the same capabilities can increase the need for automated security testing, monitoring and response. Nvidia wants to supply the full technical stack behind that shift—from accelerated computing and networking to open models and industry-specific agentic systems. 219

For buyers, the practical question is not whether AI will matter in cybersecurity. It is whether a proposed tool has the specialized models, security data, human oversight and integration into real response workflows needed to make its automation reliable. Nvidia and CrowdStrike’s SafeMind effort is an early, concrete example of that full-stack approach. 1922