SaaS is not dead, and anyone claiming otherwise, in our view, doesn’t understand what enterprise SaaS actually is, or what AI has evolved into. Before going further, a disclosure that Blossom Street is a venture fund that invests in classic B2B enterprise SaaS, so there is bias here. That said, we’re in good company. Nvidia Chief Executive Jensen Huang weighed in: “There’s a whole bunch of software companies whose stock prices are under a lot of pressure because somehow AI is going to replace them,” Huang said late Tuesday at an event hosted by Cisco . “It is the most illogical thing in the world.”
The confusion starts with a category error. AI, in its current form, creates things like content, code, and point productivity tools. That’s valuable. Enterprise SaaS does something entirely different. It manages data, orchestrates workflows, enforces protocols, and integrates across dozens of systems all while maintaining security and compliance. It becomes the operational backbone of large organizations. So AI cannot replace enterprise SaaS because it’s evolved into something fundamentally different.
Some argue that AI can now build enterprise SaaS. That is wrong. Enterprise SaaS is not just code. Companies don’t spend $50k+ a year on lines of code. They spend it on solutions that solve their problems and continue to become more innovative and robust. Solving problems means the enterprise SaaS product comes with a CS teams that ensures adoption, product and engineering teams that continuously improve the offering, and leadership teams that decide where the product goes next. Enterprises aren’t ripping out mission-critical software for something someone vibe-coded in a back room.
The argument we’ve seen more recently for AI replacing enterprise SaaS is that “companies will just build internal tools.” That’s not new. Spinning up software has always been easy. For instance Facebook built a Snapchat clone in 14 days. The initial version of Twitter was built inside of a week. The hard part isn’t writing code, it’s running the companies that sell, manage, maintain, and improve the products. Enterprise SaaS has always been threatened by home-made, internal tools, and that threat never knocks off great enterprise products. It is very hard to build, maintain, and improve enterprise SaaS products, so the idea of an internal IT team having the knowhow to not only build their own versions of Salesforce or Adobe, but also improve the products, maintain them, evolve them with new features, onboard new employees onto it, etc is absurd.
If anything, enterprise SaaS stands to benefit the most from AI. Incumbent SaaS companies already have customers, distribution, and deep knowledge of real business problems. Series A+ companies have the added advantage of speed. If you think SaaS leaders are sitting still while AI evolves, you’re mistaken.
One last question that isn’t getting much airtime: why are frontier AI labs shipping basic productivity tools at all? Anthropic’s legal product is impressive but it looks a lot like productivity software, not a replacement for enterprise infrastructure. Meta recently cited a 30% increase in output per engineer from AI coding tools. That’s meaningful but it reinforces the point. AI is boosting productivity, not replacing the backbone of the enterprise.
To SaaS founders: keep building. The market is emotional, the narrative has overshot reality, and enterprise SaaS will be the real beneficiary of the AI wave so long as AI can fulfill it’s promise.
Thank you for your readership. See more blogs and SaaS data at blossomstreetventures.com. Email the author at sammy@blossomstreetventures.com.