Ellison: AI Won't Kill SaaS, It Will Finally 'Finish' Apps

Ellison's counterpunch to the SaaSpocalypse

The Oracle co-founder used his AI World keynote to reject the idea that businesses will simply build their own software with AI and abandon commercial SaaS.

"There's been some talk about there's going to be a SaaSpocalypse because everyone's going to use AI to build their own applications," Ellison said on stage. "I don't think that's quite true. I think companies will continue to build applications, but they will be quite different because of AI."

It's a notable stance from someone whose company sells exactly the kind of enterprise software that AI is supposedly about to make redundant. But Ellison's argument isn't defensive — it's that today's apps are simply unfinished. "They may be the best things that are commercially available... [but] they're not done," he said, pointing to the extensions, connectors and third-party tools businesses still need to stitch together.

His pitch: AI gives developers the leverage to complete those applications, one industry at a time. "For the first time... [the] software industry can actually finish the applications if we start doing them industry by industry." Oracle's most visible push is healthcare, a sector Ellison described as notoriously complex and heavily regulated — exactly the kind of vertical where a generic global app struggles.

The interface disappears, the data stays

Ellison also sketched a future where workers don't learn software at all — they just ask for what they want.

"If you can ask a question describing what you want to see, you should then immediately see it. Any other user interface is nonsense," he said, adding that "there should be no training required to use any of these applications."

That's a direct challenge to the dashboard-and-menu design that has defined enterprise software for decades. And Oracle's own executives are already preparing for it. In separate interviews with TechRadar Pro, NetSuite COO Brian Chess said customers might eventually reach NetSuite through AI assistants like ChatGPT or platforms like Slack — and that's fine. "If they never log in, that's ok with us," Chess said.

The implication for software vendors is uncomfortable but clear: the value shifts from the polished interface to the data, logic and processes underneath. If Oracle remains a company's system of record, it doesn't matter whether employees touch Oracle's UI or ask an AI assistant to do it for them.

What the market made of it

Investors didn't treat Ellison's keynote as a warning sign.

Oracle shares climbed roughly 14% to $144.14 over the couple of days of Oracle AI World, according to TechRadar — a lift likely helped by both Ellison's software thesis and the company's broader product announcements. Context matters, though: the stock is still well below its late-2025 high of $330.34.

The SaaSpocalypse fears aren't fully put to bed by a keynote. But Ellison's framing reframes the question. The risk isn't that AI erases enterprise software; it's that AI makes the interface — the part users see — far less valuable than the workflow and data behind it. For vendors, that's a different kind of pressure.

What it means for you

For US, UK and EU businesses, Ellison's comments are less about a product launch and more about procurement strategy. If Oracle's vision holds, the enterprise apps you buy in the next few years will be judged on the completeness of their industry workflows and the quality of their data model — not on how nice the dashboard looks.

Oracle hasn't announced specific pricing or availability changes for AI-native versions of its applications. The NetSuite comments suggest access through third-party AI assistants may become more common, but no timeline has been given.

Ellison's keynote won't settle the SaaS debate, but it draws a clearer line: AI is coming for the interface, not the system of record. For software buyers, that means the next vendor evaluation should focus less on what the screen looks like and more on what the platform actually knows.