In early May, Confluent released a new cluster mode entirely based on S3 called the [Freight Cluster](https://www.confluent.io/blog/introducing-confluent-cloud-freight-clusters/). This is certainly not good news for Confluent's challengers.<p>Warpstream: Most affected. Confluent has the first-mover advantage and simultaneously offers low-latency GP clusters as well as the Freight Cluster, which sacrifices latency for resilience and cost benefits. If Warpstream does not offer innovations that are significantly stronger than the Freight Cluster, it will be difficult to attract those users who are less sensitive to latency from Confluent.<p>Redpanda: A Kafka-compatible streaming system rewritten in C++. It utilizes a Thread per core architecture and a no JVM setup to achieve extremely low latency. For users who are focused on ultimate latency performance and prefer a no JVM Kafka system, the impact will be minimal.<p>AutoMQ: Forked Kafka's code and redesigned the storage layer to integrate both EBS and S3, without additional latency sacrifices when using S3 as the primary storage. In terms of storage architecture, it can simultaneously offer the advantages of both Confluent's GP cluster and Freight Cluster. Since it can still ensure low latency on top of low-cost, stateless brokers, it maintains certain advantages over Confluent's storage solutions and thus will not be affected by this release.