The "AI kills software" narrative is wrong. $NOW will be the proof.

The bear case assumes enterprises will rip out their software stacks and build AI agents from scratch. That's not how large organizations work, and the most credible voice in AI agrees.
The deeper argument here isn't about ServiceNow specifically, it's about how large organizations actually work. A Fortune 500 company doesn't wake up one morning and decide to rip out its entire software infrastructure to chase an AI trend. These systems are embedded in compliance frameworks, procurement processes, and decades of institutional muscle memory. ServiceNow is already inside. SAP is already inside. Salesforce is already inside. That's not a weakness, that's the moat. When AI agents become standard, enterprises won't build them from scratch. They'll get them from the provider they already trust, already pay, and already depend on. The risk is real, a competitor could move faster, train better models, price more aggressively. But the distribution advantage of an entrenched software provider is enormous and chronically underestimated. Even startups, who might experiment with building their own agent stacks today, eventually grow into enterprises that need governance, compliance, and reliability, and they'll end up at Salesforce or ServiceNow anyway. Software as a service was already a massive industry. AI delivered through software as a service is the same industry, repriced upward, with stronger lock in. Every company will benefit from using their existing provider to access AI. Most of them just don't know it yet, but they will.
Jensen Huang put it plainly: the markets got it wrong.
-
AI agents won't cannibalize enterprise software. They'll use it more. He didn't just say it, he showed up at ServiceNow's Knowledge conference for the third consecutive year, because Nvidia runs on ServiceNow. He's not a partner endorsing a partner. He's more a customer testifying to the thesis.
-
The thesis: service is now software, and the service industry is 100x larger than the software industry.
The numbers back it up. Subscription revenue +22% YoY. Now assist customers spending over $1m in ACV grew 130%+ YoY. AI revenue set to blow past $1.5B this year. $30B targeted by 2030, double what the entire company generates today.
-
So, my take is that AI isn't replacing enterprise software. It's being delivered through it, at higher prices, with deeper lock in and a wider moat.
Huang said it best at GTC: “Is SQL going to die because agents are here? No. That’s where our ground truth of business is going to be.”
-
The market is pricing in disruption. The CEO of Nvidia is on stage saying the opposite. The numbers agree with Huang. I agree with him too, and that's exactly why I'm invested in $NOW