Nokia and Databricks launch a unified, cloud-agnostic data platform to scale real-time AI and autonomous operations across telecom networks without rewriting code.
Nokia and Databricks have successfully completed a joint proof of concept (PoC) demonstrating a unified, substrate-agnostic data platform designed for AI-driven autonomous networks. Led by Oguz Sunay (Nokia) and Nevash Pillay (Databricks), the initiative helps telecom operators resolve data fragmentation across hundreds of legacy business and operational support systems.
Simulating a performance management use case at tier-1 cloud scale, the engineering teams achieved major technical milestones. They developed cross-platform data pipelines that run seamlessly across Databricks and open-source stacks (Apache Flink, Kafka, and Iceberg) without modifying code. To prevent vendor lock-in, Nokia engineers authored the platform’s core transformation logic using an abstract expression in Python.
A custom compiler automatically translates this abstract logic on the fly during deployment into environment-native formats like Delta Live Tables or Flink SQL. Additionally, an intelligent data fabric agent can ingest natural language prompts to generate new data products, request human validation, and auto-deploy pipelines.
This specialized data fabric is tailored for multi-agent AI ecosystems, optimized for high-speed query-time data processing, zero-copy sharing, and selective upper-layer cloud ingestion. By unifying these siloed environments, the architecture enables network operators to deploy real-time analytics and scale AI operations consistently across diverse cloud frameworks without expensive software rewrites.
Moving forward, Nokia and Databricks plan to expand these autonomous network capabilities, transitioning telecommunication providers into a future where AI applications can seamlessly access, correlate, and act on large-scale network data in real time.
