

The Rise of the Context Company: Reshaping Data Engineering with Saket Saurabh
11 snips Oct 1, 2025
Saket Saurabh, CEO of Nexla and a force in data engineering, delves into the transformative shift in AI from model-focused companies to those leveraging unique contextual data. He discusses the looming limits of current AI models, the vital role of knowledge graphs for reliable AI outputs, and how tools like Lovable and Cursor are redefining workflows for developers. The conversation also highlights the evolving landscape of data engineering, where AI is now a primary consumer, necessitating new approaches to architecture and semantics.
AI Snips
Chapters
Books
Transcript
Episode notes
Model Companies Versus Context Companies
- Companies will bifurcate into 'model companies' that build foundational models and 'context companies' that own unique data and integrations.
- Owning context (documents, apps, integrations) becomes the durable enterprise advantage, not model training alone.
Limits Of Transformer Models For Processes
- Transformer-based generative models may excel at generation but hit limits for reliable, complex business processes.
- Error rates and probabilistic outputs make scaling agentic automation for critical processes risky without deterministic layers.
Early Streaming Startup Failed By Timing
- Saket recalled starting an online video streaming company before the 2000 crash that failed due to timing and technology.
- He uses that experience to argue new tech waves can be right idea but wrong timing.