Jensen Huang: AI Agents Will Use SaaS, Not Kill It
Why Nvidia’s CEO Thinks the SaaS Sell-Off Is Wrong
After weeks of panic over the “SaaSpocalypse” — a narrative where AI agents render traditional software companies obsolete — Nvidia CEO Jensen Huang went on CNBC and said the market fundamentally misread the situation. His thesis is simple and counterintuitive: AI agents won’t replace enterprise software. They’ll become its biggest power users.
The core argument: “I think the markets got it wrong.” Huang rejects the idea that agentic AI threatens companies like ServiceNow, SAP, Cadence, and Synopsys. Instead, he argues these tools exist for fundamentally good reasons — they’re systems of record, workflow engines, and operational infrastructure that agents need to do their jobs.
Agents as tool users, not replacements: “These agentic AI will be intelligent software that uses these tools on our behalf and help us be more productive.” He compares AI agents to robots entering a kitchen — they’ll use the existing microwave and the existing tools rather than reinventing them from scratch. The same logic applies to enterprise software: agents need CRM systems, ERP platforms, and analytics tools to complete their work and store results in ways humans can understand.
Nvidia’s own evidence: Huang points to what’s happening inside Nvidia itself. The company has 42,000 human employees and is rapidly deploying “hundreds of thousands of digital employees.” The result? The number of C compilers and Python programs being used is growing fast — because agents consume more software tools, not fewer. “Nobody’s going to service better than ServiceNow, and they’re going to come up with agents that are really fine-tuned and optimized for the work that uses the tools that they have.”
Market context: The IGV software ETF dropped nearly 30% in early 2026 as investors fled traditional software names — Salesforce, ServiceNow, SAP, Oracle — on fears that AI agents would make their products irrelevant. Huang’s comments came hours after Nvidia posted earnings that beat expectations, driven by sustained AI infrastructure demand.
What This Means for the AI Agent Economy
- Agents increase software consumption — More digital workers means more tool usage, not less. The installed base of “users” grows dramatically when agents join the workforce.
- Systems of record survive — Agents need structured data stores and workflow engines. They can’t operate in a vacuum — they need to read from and write to existing business systems.
- SaaS companies become AI platforms — The winners won’t be the companies that fight agents but the ones that make their software agent-accessible via APIs, MCP integrations, and consumption-based pricing.
- The sell-off may be overdone — If Huang is right, software stocks priced for extinction are mispriced. The bear case assumed replacement; the bull case is amplification.
The Counter-Narrative to the SaaSpocalypse
Huang’s argument is the most prominent counter to the “SaaS is dead” narrative pushed by leaders like Klarna CEO Sebastian Siemiatkowski, who cut 50% of headcount and predicted AI would destroy switching costs that protect incumbent software. The truth likely sits between these poles: some SaaS categories face real displacement (simple CRUD tools, basic automation), while systems of record and deep vertical software gain value as agent adoption increases. The question isn’t whether agents will use tools — it’s which tools will be worth using.