IBM and Confluent integrate Granite Time Series models for real-time inference

IBM and Confluent have integrated IBM's Granite Time Series Foundation Models (TSFM) into Confluent Cloud, enabling real-time forecasting and anomaly detection directly within Apache Flink.

• The integration allows users to call models via AI_FORECAST and AI_DETECT_ANOMALIES Flink SQL functions without managing separate ML infrastructure.
• A portfolio of four complementary models is available: PatchTST-FM for distribution forecasting, FlowState for continuous dynamics, TTM for high-volume CPU inference, and TSPulse for multi-task anomaly detection.
• Confluent Cloud provides native inference with zero configuration, handling model serving, scaling, and governance while keeping data within the platform.
• The solution supports stream-native processing, allowing models to act on live business signals rather than stale batch data.

This partnership enables domain experts to deploy time series intelligence quickly, reducing the time between an event and actionable insight from days to seconds.