World Congress 2025

Next-Gen Vector Database for Sustainable GenAI

July 9, 2025 16:00 – 18:00 Β· 120 min M1 (60 Seats)

What this session covers

Monolithic databases waste energy, powering unused data in GenAI workloads. According to the 80/20 rule, only up to 20% of the data is regularly needed, while 80% is unused. The Eclipse open-source community works on the future of vector databases: a micro vector database network - inspired by the serverless microservice architecture principle. Unlike clunky, monolithic databases that waste energy running massive servers 24/7, whether data is accessed or not, a micro vector database network dynamically allocates compute power only to the 20% of data in use, shutting down up to 80% of idle regions. The result is a blazing-fast vector search for cutting-edge GenAI solutions, slashing energy, CO2 emissions, and infrastructure costs by up to 80%. Learn how this Eclipse project redefines performance and sustainability due to the Eclipse OpenJ9 JVM, aiming to globally reduce GenAI’s energy footprint with scalable, impactful innovation.

World Congress 2025

Related talks

Open session

How to Decipher User Uncertainty with GenAI and Vector Search

  • Ben Greenberg

    Arbitrum

    Senior Developer Relations Engineer

Open session

Beyond AI Readiness: Scaling Data-Driven Engineering for Next-Gen Software & Infrastructure

  • Eduard

    Finanz Informatik

    Data Engineer Data Analytics Platform

  • Jannis Eickenroth

    Finanz Informatik

    Head of Analytics and Customer Engagement

  • Sophie Kleinke

    Finanz Informatik

    Data Engineer

Open session

Harnessing the Power of Open Source's Newest Technologies

  • Francisco Carneiro

    Eclipse Foundation

    Ecosystem Development Manager

Open session

Beyond GPT: Building Unified GenAI Platforms for the Enterprise of Tomorrow

  • Kapil Gupta

    E.ON Digital Technology

    Chief Data Technologist

All sessions at this congress