NEWS · MODELS · #3837
IBM time-series foundation models run natively on Confluent for real-time forecasting and anomaly detection
IBM’s family of time-series foundation models — with over 50 million downloads — are now available as built-in functions in Confluent (Cloud and on‑prem via Confluent Platform for Apache Flink), enabling streaming SQL to perform real‑time forecasting, anomaly detection and related tasks; the models are compact enough to run on CPUs. Confluent positions these models inside its Streamhouse architecture so Flink pipelines can invoke predictions and serve results alongside event streams.
KEY POINTS
- IBM’s family of time-series foundation models — with over 50 million downloads — are now available as built-in functions in Confluent (Cloud and on‑prem via Confluent Platform for Apache Flink), enabling streaming SQL to perform real‑time forecasting, anomaly detection and related tasks; the models are compact enough to run on CPUs.
- Confluent positions these models inside its Streamhouse architecture so Flink pipelines can invoke predictions and serve results alongside event streams.
- Bringing pretrained time‑series foundation models directly into a streaming platform reduces deployment friction and lets operational applications act on forecasts and anomalies with lower latency and without GPU dependency.
WHY IT MATTERS
Bringing pretrained time‑series foundation models directly into a streaming platform reduces deployment friction and lets operational applications act on forecasts and anomalies with lower latency and without GPU dependency.