At Norges Bank Investment Management (NBIM), which manages Norway's $2 trillion sovereign wealth fund, data timeliness isn't an abstract goal—it's a daily necessity for portfolio managers across Oslo, New York, and Singapore. With over 150 data developers working on dozens of decentralized projects, their platform began showing strain. The solution, implemented over three months, wasn't a wholesale overhaul but a targeted adoption of dbt's Fusion engine and its State-Aware Orchestration.

Senior Data Engineer Øyvind Barsnes Eraker explains the initial driver was developer experience, particularly for business-oriented teams using AI coding tools. "We needed to bridge a gap," he says. The team started deliberately, applying Fusion to five new, well-structured projects rather than legacy systems. These included platform metadata tracking, communications insights, and consolidated investment data from external vendors.

The results shifted the project's trajectory. Parsing accelerated, linting improved, and end-to-end pipeline runs consistently finished 30 to 40 percent faster without specific optimization work. The state-aware orchestration allowed models to run more frequently without increasing overall runtime. For Eraker, this performance gain was a welcome secondary benefit to the improved developer workflow.

The implications are significant for NBIM's most demanding pipelines. Teams in Singapore currently run batch processes to ensure data is ready for global trading desks, operating under tight service-level agreements. Fusion opens a path away from nightly batches toward more continuous delivery, reducing the risk of missing a critical deadline. What began as a tooling upgrade for developers has become a strategic lever for global data delivery, proving that efficiency gains often start by empowering the people writing the code.

Source: dbt Labs Blog