Keep pulling the thread on Gustav Söderström.
Unlike traditional software with near-zero marginal costs, AI systems have a high marginal cost for inference that must be covered by the business model, likely through subscriptions or tiered pricing.
Spotify achieved profitability for the first time since its founding in the current year.
According to music labels, Spotify has twice the user engagement and half the churn rate of competing music services.
Spotify's leadership views its earlier strategy of pursuing exclusive podcast content as a 'bad bet' that was abandoned in favor of a platform approach with a maximal catalog.
Spotify's leadership believes that for companies, adopting AI is an existential imperative, similar to the previous technological shifts to smartphones and the internet.
The Model Context Protocol (MCP) is an emerging industry standard that enables natural language interaction with a company's internal services and infrastructure.
A major challenge for large companies like Spotify in adopting AI is the extensive 'old school engineering' work required to expose 15 years of legacy data via real-time APIs for reasoning engines to use.
Spotify has always operated with a marginal cost business model due to royalty payments to music labels, making it well-adapted to the high marginal costs associated with AI inference.
Spotify pays out approximately 70% of its revenue to the music industry and has historically reinvested the remaining 30%, plus additional capital, back into the industry, operating unprofitably for 15 years.
Spotify is approaching 300 million paid subscribers and 700 million monthly active users.
Spotify's pivot away from exclusive podcast deals saved the company significant costs and contributed to its current profitability and improved catalog.
The widespread adoption of smartphones created an existential crisis for Spotify's early business model, which relied on a desktop-based free tier to acquire users.