As views coexist over whether AI will upend the software-as-a-service industry in a so-called “SaaSpocalypse” scenario, LinkedIn founder Reid Hoffman (리드 호프먼) weighed in.

His view can be summed up as this: the idea that AI can instantly build SaaS is far from reality, but it is true that AI is lowering the barriers to entering existing SaaS businesses.

In a post he recently shared on social media platform X (Twitter), he wrote: “The idea that you can whip up an HR system, an accounts receivable platform or an enterprise CRM with a few lines of prompts is far from reality. Even setting aside the security and compliance issues that come with directly generating code to build a company payroll system, software is not a bundle of code you build once and finish. It is a living system that requires maintenance, validation, security, regulatory compliance and continuous improvement.” He also made clear that “the existing SaaS model is being shaken” is a completely different argument from “no one will sell software.”

Even so, Hoffman also agrees that AI is bringing significant change to SaaS business models and that it has become more likely that SaaS companies will struggle to survive by doing things the way they always have.

“It is true that it will be difficult for the previous SaaS model to remain intact,” he said. “In the past, products that met customer needs were rare, and companies could secure 40 to 50 percent margins on the back of high reliability. Now, thanks to AI, the barrier to entry represented by engineering talent has fallen. It is true that the moat around SaaS has become shallower.”

He said the essence of AI-driven change in SaaS is that software must now embrace AI as a core function to stay competitive.

“The new moat comes from how precisely you have integrated AI into each field,” he said. “A company that builds a CRM agent that improves sales flows on its own, understands the pipeline more deeply than any human analyst, and has powerful back-end libraries specialized to that domain has secured a solid moat. Incumbent leaders who understand this will survive. Companies that cannot do this will be淘汰, but they will disappear more slowly than most people expect.”

Hoffman also said AI could expand the size of the software market.

“When we shifted from on-premise to cloud SaaS, the world did not end. The market actually grew,” he said. “The shift from cloud to AI-native is the same. Jevons' Paradox still works. If software development costs fall sharply, demand for software surges.”

He said change in revenue models is inevitable. “A model could spread where you pre-load a token budget and use it, like an electricity bill,” he said. “Think about a CRM company reshaping its billing system based on computing usage and scale.”

While significant changes to business models are expected, he said traditional moats such as network effects, customer relationships and data advantages will still carry weight in SaaS. “As AI becomes able to actually use proprietary data, the value of rare data rises,” he said. “Companies whose AI systems have learned years of work flows enjoy a stronger customer lock-in effect.”