If you’ve been watching the markets this week, you might think the software industry just received a terminal diagnosis. In a sudden, shivering wave of panic—now dubbed the “SaaSpocalypse”—global software stocks took a massive hit, wiping out over $250 billion in market value in a matter of days.
The catalyst? A new suite of automated plugins from AI startup Anthropic. But while investors were hitting the “sell” button, Nvidia CEO Jensen Huang was doing something else: shaking his head.
Speaking at a Cisco-hosted conference in San Francisco, Huang didn’t mince words. He called the panic-selling of software stocks “the most illogical thing in the world.” —
The Panic: What Triggered the “SaaSpocalypse”?
The selloff wasn’t a random event. It was sparked by the release of Claude Cowork, an AI agentic tool that can automate complex tasks in professional services—specifically legal work, contract reviews, and compliance.
When the market saw an AI that could “do the job” of traditional software and specialized services, the reaction was swift:
- Infosys: Plunged 7.3%
- Thomson Reuters: Dropped 15%
- LegalZoom: Declined nearly 20%
- Global Impact: From India to Japan to the U.S., software valuations cratered as the fear of “AI replacement” went viral.
Jensen’s Logic: AI Uses Tools, It Doesn’t Kill Them
Jensen Huang’s rebuttal is built on a simple, grounded truth about how technology actually works. According to Huang, the idea that AI will replace software tools is a fundamental misunderstanding of AI agency.
“If you were a human or a robot, or an artificial intelligence, would you use existing tools or reinvent them? The answer is obviously to use existing tools.” — Jensen Huang
Why Software is Safe (For Now):
- AI Needs “Explicit” Tools: AI models are great at reasoning, but they aren’t great at being a spreadsheet or a legal database. They need “explicit” software tools to execute their plans accurately.
- The “Tool-Use” Breakthrough: The latest advancements in AI are focused on tool usage. AI is becoming a better user of software, not a replacement for it.
- Physical AI: As we move toward 2026’s biggest trend—Physical AI—AI needs to interact with the real world and existing digital infrastructure. You don’t delete the operating system just because you have a smarter driver.
A “No-Win” Situation for Nvidia?
Ironically, while Huang is defending the software industry, Nvidia itself is in a strange spot. Despite record earnings and the full production of the new Vera Rubin superchip, Nvidia’s stock has faced its own volatility.
Huang recently admitted to employees that the company is in a “no-win” situation regarding market sentiment:
- If they miss a target, it’s an AI bubble.
- If they beat a target, they are fueling the bubble.
By defending software, Huang is essentially defending the very ecosystem that justifies the massive spending on his chips. If software dies, the demand for AI hardware eventually follows.
The Bottom Line: Evolution vs. Extinction
The market has a habit of overcorrecting. In 1999, people thought the internet would kill every physical store. It didn’t; it just changed how they operated.
Jensen Huang is betting that the same applies here. AI isn’t the “Software Killer”—it’s the Software Power-User. The companies that will thrive aren’t the ones trying to out-reason the AI, but the ones building the high-quality tools that AI agents can’t live without.
SEO Quick Stats: The Impact
| Sector | Market Cap Loss (Estimated) | Biggest Losers |
| Legal Tech | $45 Billion | LegalZoom, RELX |
| IT Services | $90 Billion | Infosys, Wipro |
| Enterprise SaaS | $115 Billion | Salesforce, SAP |
What’s your take? Is the market right to be scared of Anthropic’s automation, or is Jensen Huang right that this is just another “illogical” panic?
Would you like me to analyze which specific software sectors are actually the most “at-risk” versus those that are safe “tools” for AI?
Leave a Reply