EDB and NVIDIA Fuel the Rise of the Agentic AI Workforce

WILLMINGTON, Del. – March 16, 2026 – EnterpriseDB (EDB), a company specializing in the Postgres database, today announced a major collaboration with NVIDIA that promises to redefine the data infrastructure for enterprise artificial intelligence. By integrating NVIDIA's AI infrastructure, EDB is accelerating its Postgres platform to deliver performance improvements of up to 100x for interactive analytics, aiming to equip businesses for what it calls the "agentic workforce era." This move positions the popular open-source database as a critical foundation for a new generation of autonomous AI systems that can reason, collaborate, and take independent action across vast corporate datasets.

The Dawn of the Agentic Workforce

The term "agentic workforce" signals a significant evolution beyond the current wave of generative AI, which primarily responds to human prompts. This new paradigm involves coordinated systems of autonomous AI agents that can independently set goals, create plans, and execute complex tasks with minimal human intervention. These digital workers are designed to handle entire workflows, from data analysis and report generation to orchestrating supply chain logistics, fundamentally changing the nature of work. According to industry analysts at IDC, 45% of organizations are expected to orchestrate AI agents at scale across business functions by 2030.

However, the transition to this new era is fraught with technical challenges. Most organizations are hampered by data that is locked away in separate, siloed systems. This "data ping-pong," as IDC research director Devin Pratt describes it, creates inefficient analytics pipelines and stalls progress. Agentic systems require real-time access to live transactional data, historical context, and analytical datasets simultaneously—a capability most current architectures cannot support without significant latency and risk. "The arrival of the agentic workforce demands a rethink of data architecture," Pratt noted in a statement. EDB's new offering aims to be the high-velocity foundation needed to operate these systems at scale.

A Sovereign Foundation for Enterprise AI

As AI agents become more autonomous and embedded in critical business processes, concerns around data governance, security, and compliance have escalated. In response, the concept of "Sovereign AI" has become a top priority for enterprises, particularly in regulated industries like finance and healthcare. Sovereign AI refers to an organization's ability to deploy and govern AI systems using its own controlled infrastructure, ensuring that sensitive data remains within its legal and strategic boundaries. Research from EDB indicates that over 90% of global enterprises see building a sovereign AI platform as a mission-critical priority within the next three years.

EDB is directly addressing this need with its EDB Postgres AI (EDB PG AI) platform, which is engineered to be "sovereign by design." The platform enables organizations to deploy powerful AI workloads while maintaining complete control, offering fully air-gapped support for environments with no external internet access. This allows companies to import AI models and containers into their own private registries, ensuring that proprietary data and custom models never leave their secure perimeter. This capability is crucial for moving from isolated AI experiments to production systems where autonomous agents operate across billions of sensitive records in real time.

Under the Hood: GPU-Accelerated Postgres

The dramatic performance gains announced by EDB are powered by a deep integration with NVIDIA's AI stack. The core of the acceleration comes from NVIDIA cuDF for Apache Spark, a GPU-accelerated library that allows for massive parallel processing of data. By offloading heavy analytical workloads from CPUs to GPUs, the EDB PG AI Analytics Engine can query and synthesize terabytes of data in seconds, not hours. This enables real-time conversational analytics, immediate decision-making, and complex multi-agent orchestration without the need for costly data duplication across separate data warehouses or lakes.

The integration extends beyond raw processing speed. EDB PG AI also incorporates NVIDIA NIM model serving, a set of microservices that optimize the on-premise deployment of large language models like NVIDIA's Nemotron. This is vital for agentic systems, which rely on these models for reasoning and planning. Furthermore, the platform uses NVIDIA NeMo Retriever to accelerate Retrieval-Augmented Generation (RAG) pipelines. This allows AI agents to quickly and accurately find relevant context from live enterprise data stored in Postgres, dramatically improving the quality and reliability of their actions. This combination of GPU-accelerated analytics, optimized model serving, and high-speed RAG provides a unified, high-performance environment for building and scaling sophisticated AI applications.

From Pilot to Production: Bridging the AI Chasm

While the promise of AI is vast, many enterprises struggle to move beyond small-scale pilots and experiments. EDB's own "Sovereignty Matters" research highlights this challenge, revealing that only 13% of enterprises have successfully deployed production-scale agentic systems with more than ten active workflows. However, this elite group is reaping substantial rewards, generating five times higher ROI than their peers and operating with double the density of AI agents per business process.

Closing this gap between experimentation and scalable production is the central business proposition of the EDB and NVIDIA collaboration. The platform is designed not just for speed, but for the predictability and control necessary for mission-critical systems. "Enterprises want GPU acceleration, but they also need predictability and control," said Quais Taraki, CTO at EDB. "That is the difference between impressive demos and durable production systems." By providing GPU-based workload isolation to protect query performance and integrating governance capabilities via Apache Iceberg, the platform aims to give enterprises the confidence to deploy agentic workforces at scale.

To mark the announcement, EDB is distributing complimentary copies of its O'Reilly Media book, Building a Data and AI Platform with PostgreSQL, to all 25,000 attendees at the NVIDIA GTC 2026 conference. This strategic alignment positions the open-source database at the center of the next wave of enterprise AI adoption, promising a future where intelligent, autonomous systems are not just an experiment, but a core driver of business operations.