IBM Granite Time Series Models Launch in Early Access on Confluent Cloud
IBM and Confluent announced that IBM Granite Time Series models are live in Early Access on Confluent Cloud, callable directly from Apache Flink via SQL functions for streaming forecasting and anomaly detection, with Confluent Platform to follow.
What happened
Event details
IBM and Confluent announced that IBM Granite Time Series models are live in Early Access on Confluent Cloud, with Confluent Platform to follow. The portfolio includes four complementary time series foundation models covering forecasting, anomaly detection, similarity search, classification, gap-filling, and optimization. The models run stream-native within Apache Flink on Confluent Cloud and can be called directly through Flink SQL functions such as AI_FORECAST and AI_DETECT_ANOMALIES, without requiring dedicated model-serving infrastructure or GPUs. According to the official article, the offering provides zero configuration, built-in governance, and traceability, with inference results written to Kafka topics and shareable to alerting systems, dashboards, lakehouses, and AI agents. Access opens on Confluent Cloud on AWS. IBM reports more than 44 million downloads of its time series models and says design partner deployments delivered productivity gains of 5 to 10×.
Assessment
Why it matters
This is a substantive product integration launch between IBM's Granite Time Series foundation models and Confluent's data streaming platform, enabling enterprise users to perform streaming forecasting and anomaly detection directly where data moves.
Availability
Access notes
The article describes Early Access availability, with access opening first on Confluent Cloud on AWS; Confluent Platform will follow, but no specific timing is stated.