Databricks drove down AI coding spend 70%

Databricks reduced AI coding costs by 70% by implementing cost management techniques that balance broad employee access to AI tools with budget constraints. The company identified that the efficiency frontier—models offering the best price-for-quality ratio for typical coding tasks—is advancing rapidly, making it possible to shift to cheaper, newer models without sacrificing performance. Databricks has open-sourced key infrastructure components like its AI Gateway to help other enterprises achieve similar cost reductions while maintaining AI productivity gains.