Open table formats such as Apache Iceberg or Delta Lake have transformed the data landscape. For the first time, we’re seeing a real open storage ecosystem emerging across database vendors. So far, open table formats have found little adoption powering low-latency, high-concurrency analytics use-cases. Data stored in open formats often gets transformed and ingested into closed systems for serving. The reason for this is simple: most modern query engines don’t properly support these workloads. In this talk we take a look under the hood of Firebolt and dive into the work we’re doing to support low-latency and high concurrency on Iceberg: caching of data and metadata, adaptive object storage reads, subresult reuse, and multi-dimensional scaling. After this session, you will know how you can build low-latency data applications on top of Iceberg. You’ll also have a deep understanding of what it takes for modern high-performance query engines to do well on these workloads.