Software stocks are getting destroyed. ServiceNow closed at $101 this week, down ~50% from $211 a year ago. Their Q4 2025 results? Subscription revenue up 21% year-over-year, beating guidance and key expectations. Salesforce was down 40% year-over-year. Adobe down 38%. Palantir down 15% in 2026 alone, despite beating earnings.
It seems the market isn’t reacting to performance, it is reacting to fear of AI.
Just like in politics, we’re living through a binary moment. Some believe SaaS is completely dead. Others believe it will remain essential. Rather than operating from extremes, I wanted to share what’s top of mind based on working with enterprise businesses ($1B+ rev.), not your cousin’s startup doing $2K a month.
Software as a Service is paying for platforms, so you don’t have to build or manage them fully in-house. Enterprises use SaaS for accounting (QuickBooks), marketing automation (Braze), commerce (Salesforce Commerce Cloud), loyalty programs (Kobie), and operations (ServiceNow). You login, use it, and their teams manage the infrastructure, security, numerous updates, and hosting for a recurring fee.
As a reminder, OpenAI is SaaS. Anthropic is also SaaS. You pay for the platform and usage (e.g, API).
What Happens If SaaS Actually Dies?
If SaaS is truly dying, enterprise companies will either (1) build these capabilities in-house or (2) buy the software (or get it for free) and manage it themselves ongoing.
If new AI-powered options flood the market, that’s not the death of SaaS, that’s fragmentation of the market. Paying monthly or annually for technology usage is still SaaS, even if there are thousands of new vendors to choose from.
At times it seems we’re confusing “SaaS disruption” with “SaaS extinction.”
Retailers need to create and market incredible products in a hyper competitive market. Financial services firms need to solve complex market and investment challenges. Brands must focus on their core offering.
Trying to build and maintain an enterprise technology team to support your most critical functions like marketing automation, commerce platforms, ERP, creates massive risk and pulls focus from core business. And AI resources are NOT cheap.
The assumption that AI will make custom development trivial ignores the reality of enterprise operations. Development is 20% of the problem. Maintenance, training, adoption, operations, integration, among many others are the other 80%.
Most people outside enterprise don’t understand the absolute requirements around security and governance. Would you want your bank using a vibe-coded tech stack cobbled together by AI agents?
Regulatory bodies don’t care about innovation speed when customer data is at risk.
Enterprises operate under GDPR, CCPA, SOC 2, ISO 27001, HIPAA, and dozens of other frameworks that require documented controls, audit trails, and accountability. These aren’t suggestions. They’re legal requirements with million-dollar penalties for violations.
These compliance requirements have industry and national mandates that won’t shift easily, regardless of how advanced AI becomes.
Imagine an AI agent handling all your marketing. It triggers campaigns, sends messages, optimizes spend, but you can’t see what was sent, why, or to whom. This would be a hard sell for an enterprise brand no matter how successful it is.
Across customer purchases and transactions, marketing, and CRM, brands require accessible ledgers to review transactions, analyze behavior, and manage risk. Daily operations demand it. Legal teams need documentation for disputes. Finance needs attribution for budget allocation. Compliance needs proof systems are operating within approved parameters. Direct access to this data will live on.
Investors fear that AI agents will replace seat-based pricing models. If software no longer charges per user, revenue models collapse. That fear is real. But the timeline and scope are overestimated.
What the market is missing: enterprises aren’t going to rip out Salesforce, Workday, SAP, or ServiceNow overnight. These systems have tens of millions of dollars of integration, customization, and dependency built into them. Switching costs are enormous.
The need to improve margins is real. Per-seat pricing will need to evolve. Light SaaS tools with low switching costs, limited proprietary data, and easy-to-replicate workflows are at risk (and will be annihilated).
But enterprise SaaS, platforms managing mission-critical operations with deep integrations, compliance requirements, and regulatory oversight, isn’t going anywhere soon. The market hasn’t priced this distinction yet.
SaaS isn’t dead. It’s being stress tested. And the companies that survive will be the ones solving real enterprise problems.
The selloff reflects fear, not fundamentals. Here’s what matters:
The progress in AI over the past year has been staggering. But fear is running ahead of implementation. The market is pricing in full disruption before most enterprises have even scaled their first AI agent in production.
I’m excited to partner with brands building through this moment. The intersection of AI capability (what is can actually do) and enterprise reality (how ready each brand is today) is where the most interesting work happens.