We are already using DynamoDB Streams. Why do you think we need to use Kinesis Streams again?

DynamoDB Streams captures changes to items in a single table and delivers them to consumers with a short retention period (24 hours). It is tightly coupled to DynamoDB and is best for simple use cases such as triggering a Lambda to update a cache or send notifications. Kinesis Streams, on the other hand, is a general‑purpose, scalable data streaming service that can ingest data from multiple sources, retain records for up to 365 days, and support multiple consumer applications with independent read offsets. If you need to aggregate streams from several tables, replay data for analytics, or integrate with other AWS services like Redshift, Snowflake, or third‑party tools, Kinesis gives you that flexibility. Additionally, Kinesis offers built‑in sharding for horizontal scaling and a richer set of consumer APIs (Kinesis Client Library, Firehose). Therefore, you might choose Kinesis when you need cross‑service integration, longer retention, or higher throughput beyond what DynamoDB Streams can comfortably provide.