Databricks announced the general availability of high-QPS scaling for Databricks AI Search, allowing standard endpoints to scale to thousands of queries per second with a single parameter. Users set a target_qps on the endpoint at creation or update it on any existing endpoint via the SDK, REST API, or endpoint UI, and Databricks provisions the infrastructure to meet that target without managing replica counts, node sizes, or load balancers. Unity Catalog governance and Delta Sync remain in place, enabling the same endpoint to scale from prototype to production traffic. Three real-time production patterns requiring high-QPS scaling include search bars (e.g., product search on e-commerce, content discovery on streaming platforms, voice search on connected devices), recommendation and personalization systems (e.g., "You may also like" panels, personalized feeds), and real-time entity resolution (e.g., identity matching, deduplication, lookup against large catalogs). Indicators of needing this feature include 429 (Too Many Requests) errors in production logs, P95 latency spiking as traffic ramps, and workarounds like duplicate endpoints or client-side load balancing.