The core argument is that AI is rapidly commoditizing software development and data migration, which will drastically lower the high switching costs that protect incumbent SaaS companies. This same technological force enables massive operational efficiency, allowing companies like Klarna to significantly reduce headcount while increasing output.
High-margin SaaS businesses, which have historically traded at high multiples due to sticky customers and high switching costs, are facing severe valuation compression. As AI makes it trivial to replicate functionality and migrate data, their 'utility-like' nature will be exposed, potentially causing price-to-sales multiples to fall from 5-10x to 1-2x.
Klarna is presented as a case study in aggressively pivoting to an AI-first operating model. This involves not just adopting AI tools, but fundamentally re-architecting the company around a unified tech stack, leveraging proprietary data, and innovating on labor models to achieve unprecedented efficiency.
A counter-narrative to the idea of infinite demand for compute is presented. While consumer AI applications may drive demand, enterprise AI could act as a powerful data compression tool, making operations more efficient and potentially reducing the overall need for massive compute resources in the long run.
Keep pulling the thread on Sebastian Siemiatkowski.