Nokia and Databricks Complete Proof of Concept for Cloud‑Neutral Data Platform

The test validates a compiler-driven approach that lets operators run identical analytics across commercial and open-source stacks to speed AI-led network automation.

Overview

Nokia and Databricks said they finished a joint proof of concept on June 24, 2026 that ran the same data workflows on Databricks and an open-source stack using Apache Flink, Kafka, and Iceberg.
Engineers used a platform‑independent Python layer and a custom compiler that converts abstract transformation logic into native runtimes such as Delta Live Tables or Flink SQL so pipelines need not be rewritten for each target.
The PoC demonstrated an intelligent data‑fabric agent that can create, request human validation for, and deploy new data products from natural‑language prompts, reducing manual pipeline work.
Nokia announced an expanded deal with AWS the same day to run its Autonomous Networks Fabric on Amazon’s cloud with availability expected later this year, and Nokia shares rose about 1.9% on the news.
The project aims to break telecom data silos and avoid vendor lock‑in so tier‑1 operators can run real‑time analytics at scale, which could speed network fixes, cut integration costs, and enable broader AI-driven cross‑domain decisions.