Building the Universal AI Automation Layer ft n8n CEO Jan Oberhauser
n8n CEO Jan Oberhauser explains how dropping lead-gen goals and pivoting to AI orchestration quadrupled revenue in 8 months after 6 years of flat growth.
- n8n quadrupled total 6-year revenue in just 8 months after pivoting to AI agent orchestration and abandoning lead-gen marketing goals.
- The pivot signal: Pinecone raising ~$100M reframed itself from ‘vector database’ to ‘AI memory layer’ — Jan copied that repositioning logic for n8n.
- n8n uses a ‘fair code’ license (not OSI-approved): free to use and self-host, but nobody can commercialize or resell the code, a rule Jan set from day one to avoid the backlash Redis and others faced.
- Open source self-hosting is often more expensive than cloud, not cheaper — idle hardware running 99% of the time is the real cost; data privacy and control are the actual enterprise drivers.
- Jan sees MCP as ‘the HTTP of AI workflows’ — a good-enough standard that unlocks agent-to-agent and marketplace plug-in models regardless of which protocol wins long-term.
- He believes horizontal orchestration platforms beat vertical AI apps long-term: vertical tools multiply complexity, which then requires an orchestration layer anyway.
- AI foundation model progress is slowing (GPT-5 as signal), but Jan expects another acceleration wave as early-stage research matures and capital finds new approaches.
- Breakout category pick for next 6–12 months: AI-powered internal tooling, where organizations can take more risk than in consumer-facing products.
2025-08-26 · Watch on YouTube
