The conversation emphasizes that true defensibility and long-term value in software come from building platforms, not individual products or 'tools'. Platforms create stickiness through deep integration into customer workflows and ecosystems, making them difficult to replace, which is why so few software companies reach the $10B+ revenue scale.
A dominant thesis among investors is that AI will commoditize existing SaaS applications, potentially driving the terminal value of some companies to zero. Incumbents are now under intense pressure to prove that AI can re-accelerate their growth, not just cannibalize their existing business.
Amidst the uncertainty in the application layer, the data layer and the LLM layer are identified as two permanent, non-negotiable components of the modern AI stack. Companies that are foundational in these areas, like MongoDB for data, are positioned to be less disruptable and more essential as AI adoption grows.
Despite the market hype, large enterprise adoption of AI is proceeding cautiously and unevenly. While developer-focused tools like coding assistants are seeing rapid, successful adoption, broader applications like office productivity co-pilots and AI-native customer support are still in experimental phases with unclear value propositions for many customers.
Keep pulling the thread on CJ Desai.