Data analytics is moving into the cloud and many highly successful commercial cloud-native systems exist. However, existing systems are either expensive, lock in user data, or do both. In this talk, I will present ideas for how to commoditize large-scale data analytics in the cloud. Commoditization involves (a) reducing query processing cost to the theoretical hardware performance limits and (b) avoiding vendor lock-in by making it easy to move data between different clouds and systems. The architecture of an open and cost-efficient data analytics system can be split into three main components: First, an intelligent control component that automatically and transparently selects and manages the cheapest hardware instances for the given workload and makes migration to other cloud vendors possible. Second, a highly-efficient and scalable query processing engine that is capable of fully exploiting modern cloud hardware. Third, a data lake storage abstraction using open data formats that enables cheap storage as well as modularity and interoperability across different data systems.