Jensen Huang Says Cybersecurity Is AI’s Next Big Market

Nvidia CEO Jensen Huang’s central claim at Goldman Sachs’ Communacopia + Technology Conference was straightforward: AI’s success in software development makes cybersecurity a natural next market. If AI can write code, it can also help identify bugs, simulate attacks, validate defenses and accelerate remediation. Huang called cybersecurity likely to become the next major application of AI—and, because the work is continuous, a significant commercial opportunity. 310

Why Huang sees cybersecurity as the next AI application

Huang framed security as an extension of coding. Bug finding follows code generation, and security operations turn that process into an ongoing contest: red teams search for weaknesses while blue teams patch and defend them. As AI speeds up software creation, it can also speed up vulnerability discovery and remediation. 310

That matters because the proposed use case is not a one-time coding assistant. Huang’s thesis is a continuous operational loop of testing, detection and fixing. This is the business logic behind his prediction that cybersecurity could be a large, durable AI market rather than merely another feature inside developer tools. 610

What Nvidia’s partnerships actually show

The companies Huang mentioned should not be treated as identical cybersecurity partnerships. The available announcements point to different roles.

CrowdStrike: the most direct cybersecurity example

Nvidia and CrowdStrike unveiled SafeMind, an agentic cybersecurity system that combines Nvidia Nemotron models with CrowdStrike technology and proprietary data. CrowdStrike described it as a continuous penetration-testing loop with offensive and defensive models and a simulated digital-twin environment. 14

This is the clearest example of Huang’s argument in practice: AI agents can be used to repeatedly probe systems and improve defenses rather than simply generate software.

Cisco: an AI-security ecosystem role, but limited detail

Huang named Cisco among companies participating in the AI-security ecosystem. Reporting indicates Nvidia’s role is focused on supplying AI-factory and inference infrastructure, rather than Nvidia becoming a standalone cybersecurity vendor. However, the supplied material does not establish a specific Cisco–Nvidia cybersecurity product built on Nemotron models. 28

Palantir: sovereign AI and supply chains, not a dedicated cyber product

Nvidia and Palantir did announce a collaboration involving custom Nvidia Nemotron open models. But the announced deployment is a sovereign-AI stack for complex supply-chain operations, beginning with Nvidia’s own supply chain. It is not, based on the announcement, a dedicated cybersecurity partnership. 18

The distinction is important. Palantir may be relevant to Nvidia’s broader enterprise-agent strategy, but it should not be presented as evidence that Nvidia and Palantir launched a joint cyber-defense product.

Huang’s response to AI fear narratives

Huang pushed back on claims that advanced AI will inevitably cause human extinction, reportedly calling that idea “total nonsense.” He also rejected a simple story in which AI only destroys jobs, presenting AI instead as a technology that changes work and raises productivity. 42

He made a separate and more provocative argument about the cybersecurity discussion itself: vendors preparing new products may have an incentive to emphasize threats and intensify demand. Huang said, “what better way to create demand than to create a problem?” 45

That comment should not be read as a claim that cyber risks are imaginary. His own cybersecurity thesis depends on AI increasing both the scale of software activity and the need for continuous defense. Rather, he was challenging the commercial incentives behind some of the public messaging around AI security.

The competing safety view

The debate around AI risk remained sharply divided. Jacob Coxon, a former researcher at Anthropic and OpenAI, said he resigned because the labs were “gambling with our lives” in a race toward self-improving superintelligence. He said people building AI believed it could “kill us all by the end of the decade.” 47

Bill Gates has separately argued that governments and societies should prepare for AI-driven disruption by protecting some work for people and reconsidering institutions and tax policy. 35

These concerns are not directly refuted by Huang’s market thesis. AI can create valuable defensive tools while also raising questions about labor disruption, misuse, concentration of power and the governance of increasingly capable systems.

Huang’s AI spending and revenue outlook

Huang reiterated his forecast that global AI infrastructure spending could reach $3 trillion to $4 trillion by 2030. 5658

Reports from the conference also described Nvidia as targeting roughly 70% year-over-year revenue growth, with Huang characterizing supply—not demand—as the primary constraint. The reporting is inconsistent about the fiscal-year label, so the percentage is clearer than attaching it to a specific fiscal year without checking Nvidia’s primary filings or an official transcript. 5054

Together, the forecasts explain why Nvidia sees cybersecurity as strategically important: it expands the demand story beyond model training and code generation into continuous enterprise operations.

Huang’s answer to “circular financing” concerns

Huang also rejected concerns that Nvidia could be artificially supporting demand by investing in customers and infrastructure projects that then purchase Nvidia systems. His response was economic rather than structural: if Nvidia invests $1 and a customer creates $100 in value, further investment is rational. “Is that circular?” he said. “If that is, let’s do more of that.” 53

That is a defense of the expected value created by Nvidia’s ecosystem investments. It does not, by itself, resolve questions investors may have about the financial relationships, durability of customer demand or concentration of AI infrastructure spending.

The practical takeaway

Huang’s argument is bullish but internally consistent: AI-assisted software development increases the need for AI-assisted security, and the result could be a recurring market for vulnerability detection, red-teaming and remediation. CrowdStrike’s SafeMind offers the strongest concrete example among the partnerships discussed. 14

The broader claims require more care. Cisco’s exact cybersecurity product role is not established in the supplied reporting, and the Nvidia–Palantir announcement is primarily about sovereign AI for supply-chain operations. Meanwhile, the disagreement over AI safety and jobs remains unresolved: commercial opportunity and societal risk can exist at the same time. 183547