How Enterprises Are Building Robust Infra Backbone For AI At Scale

Enterprises in India are building a robust infrastructure backbone to support AI at scale as the country's enterprise AI market is expected to reach $71 billion by 2030. The conversation shifted from experimentation to predictability, governance, and cost control as cloud costs rise and data pipelines struggle with unstructured inputs.

Oracle India's Vivek Gupta highlighted that AI at scale is not a GPU problem but an engineering problem that needs to be solved. Ankit Mehra from GyanDhan emphasized the importance of defining clear business outcomes before building AI, as poorly framed questions often lead to more problems. HYPD's Nitish Gupta discussed the challenges posed by user-generated content on creator platforms, with varying regions and formats that make data unpredictable.

DeHaat's Sanjeev Singh echoed this sentiment, stating that the problem has evolved from a pure data challenge to a compounded data-plus-AI challenge. NimbusPost's Sanjeev Gupta highlighted that infrastructure decisions vary by industry and use case, and experimentation comes at a high cost. Pidge's Shubhanshu Chouhan cautioned against relying on opaque models in critical operations like rider and delivery allocation, as poorly governed inputs can expose sensitive information and lead to increased costs.