This article discusses how to enhance serverless computing with a GraphQL data-as-a-service layer. It explains that while serverless functions are stateless and can't hold state across invocations, they offer scalability and ease of deployment. However, this also leads to challenges in data access, such as connection pooling, performance optimization, endpoint discovery, load balancing, high availability, and security. The article suggests treating data more like a service than a database and using GraphQL for a unified data API for frontend and serverless app development. It recommends using Hasura to handle these tasks, which supports automatic mapping of database concepts into GraphQL types, fine-grained authorization, transactions over websockets, and high performance with minimal overhead.