20VC: SaaS is Dead: Why Systems of Record Will Die in an Agentic World | What Revenue Multiple Will Software Companies Trade At? | From 7,000 to 3,000: We Need Less People Than Ever with Sebastian Siemiatkowski

The conventional wisdom says AI will make software development cheaper and faster. Sebastian Siemiatkowski, CEO of Klarna, argues it will do something far more radical: kill the software industry as we know it. Not just change it or optimize it, but fundamentally eliminate the need for most of the applications businesses pay billions for today.

Siemiatkowski's thesis rests on a simple observation about switching costs. Today, companies get locked into CRM systems, accounting platforms, and productivity tools because moving data between them is expensive and painful. But what happens when AI agents can seamlessly transfer information between any system? "If the cost of software creation is going down, how do we determine which businesses have sustaining value versus which do not?" Siemiatkowski asks. His answer: "The next thing that's going to hit everyone bad is the switching cost of data. Because so far, what you're seeing is you have proprietary data stuck in, for example, the CRM vendor."

This isn't theoretical for Siemiatkowski. He spent a weekend building what he calls "company in a box" , an AI-powered workspace that handles accounting, CRM, and other business functions through connected agents rather than monolithic software platforms. The experiment convinced him that traditional SaaS companies are building the wrong thing entirely. "Why do you need to build it all yourself if you're gonna have agents that are able to move data between different products much more easily?" he argues. Instead of paying Salesforce thousands per month for a complex CRM, a plumbing company could have Claude manage customer relationships across multiple lightweight tools, switching between them as needs change.

The implications for software valuations are stark. If switching costs disappear and AI agents eliminate the need for integrated platforms, what justifies the premium multiples that SaaS companies command today? Siemiatkowski believes we're heading toward a world where most business software becomes commoditized, much like how cloud computing made physical servers obsolete. The companies that survive will be those with genuine data moats or network effects, not those that simply aggregate business functions into a single interface.

Perhaps more provocatively, Siemiatkowski argues that AI isn't just changing what software we build, but who builds it. His own experience reducing Klarna's workforce from 7,000 to 3,000 employees demonstrates how AI can eliminate entire categories of knowledge work. But he sees an even more fundamental shift coming in software development itself. "Right now, AI is allowing us to reinvent the wheel all the time," he observes, noting how multiple developers prompt AI to generate identical code fragments. The next wave will involve intelligent caching and reuse of these patterns, making software creation not just cheaper but radically more efficient.

For investors flooding money into AI startups, Siemiatkowski has blunt advice: stop investing in companies you don't understand technically. "If I meet investors today that haven't actually downloaded and tried to build something themselves," he suggests they're making decisions based on hype rather than real capabilities. He recommends VCs spend time with tools like Cursor to understand how AI actually changes the development process before writing checks.

The broader lesson extends beyond software. Siemiatkowski's vision suggests we're entering an era where the value of integration and platform lock-in collapses, replaced by fluid ecosystems of specialized tools connected by intelligent agents. Companies that built moats around switching costs and complex interfaces may find those advantages evaporating faster than they expect.

Watch for early signals in your own organization. Are your teams starting to question why they need heavyweight software for simple tasks? Are they experimenting with AI tools that connect previously siloed systems? The companies that recognize this shift early and position themselves as specialized, best-in-class tools rather than integrated platforms may be the ones that survive the coming transformation. The rest may discover that their software empires were built on foundations that AI can dissolve overnight.