This is a real-business compounding case, not a pure
rerating bet. Innodata already has demand, margins, cash generation, and low capital intensity. The upside comes from turning project revenue into a broader recurring trust layer across more customers. I assume some multiple compression from today’s AI enthusiasm, but not a collapse, because the business should be larger, less concentrated, and more software-enabled by 2031.