How a global investment firm reduced runtimes by 30–40% with the dbt Fusion engine

Norges Bank Investment Management (NBIM) manages Norway’s sovereign wealth fund, one of the largest in the world. Overseeing more than two trillion dollars in assets across global markets, NBIM holds about 1.5% of all listed companies worldwide.

To support these investments, NBIM runs a sophisticated data platform that serves investment teams across Oslo, New York, and Singapore. But as NBIM’s data practice scaled to more than 150 dbt developers across 30+ decentralized projects, the team began to feel the strain.

In this post, you’ll learn how NBIM addressed these challenges by adopting the dbt Fusion engine and State-Aware Orchestration (SAO). NBIM’s journey reveals that the best path to efficiency can start with making life easier for your developers.

A smooth rollout of the dbt Fusion engine in 3 months

When NBIM began exploring Fusion, its primary concern was improving the developer experience. SAO was viewed as a bonus.

“We’re doing a lot of work with AI agents and coding tools like Cursor,” explains ​​Øyvind Barsnes Eraker, Senior Data Engineer at NBIM. “We wanted to bridge a gap between technical data engineers and business-oriented teams who write code but may struggle with code structure.”

Rather than migrating its most complex legacy projects immediately, NBIM’s data team started with smaller projects built around fewer, well-structured models. In just a few months, NBIM set up five projects in production using Fusion and SAO. The projects focus on:

• Platform metadata: These projects track usage of dbt, Snowflake, Fivetran, and Tableau, helping the team make decisions around data deprecation and cleanup.
• Communications insights: The team can easily access PR, webpage, and social media data that informs NBIM’s communications strategy.
• Investment data: NBIM can use data from external vendors to create a single source of truth for investment teams.

“We were intentional about expanding Fusion into projects at a sustainable pace while onboarding new users,” reflects Øyvind. “It’s been a great experience with no significant hiccups at all.”

Faster runtimes with a better developer experience

By focusing on projects that can translate into quick wins, the data team has avoided migrating technical debt. Thanks to the thoughtful Fusion rollout, they’ve already experienced improvements in their workflows.

“We’re seeing better feedback loops, faster parsing times, and improved linting,” says Øyvind. “Our less technical stakeholders, like portfolio managers, are creating higher-quality projects, too.”

Importantly, since moving projects to Fusion, the team has observed faster end-to-end runs. They typically complete 30-40% faster, even without any dedicated optimization efforts.

“SAO enables us to lighten up on the orchestration so that we can run models and pipelines more often without seeing runtime increases,” says Øyvind. “We’re confident we’ll see cost optimizations downstream too.”

As the team looks ahead, they view Fusion as mission-critical for reducing SLA risk on NBIM’s most complex projects.

“Our larger pipelines still run in a very batch-oriented way,” says Øyvind. “We have teams in Singapore that run and validate the pipelines so the data is ready as early as possible for portfolio managers across different timezones. That creates very tight SLAs around when data has to be delivered.”

For these projects, there’s little room for delays. Fusion could allow the team to move away from nightly batches toward more continuous data delivery—helping the business access insights faster.

Rethinking data delivery on a global level

For NBIM’s data team, transitioning to Fusion initially was about accessing better tools. Their primary goal was to help business users write better code.

But implementing Fusion quickly evolved into an overall performance upgrade. By improving developer workflows, Fusion has enabled faster data delivery with improvements in efficiency.

"For us, the developer experience is still first and foremost,” concludes Øyvind. “But everything else on top of that—in terms of data timeliness, better testing, and better orchestrations—is a huge bonus. We expect dbt Fusion will continue to bring a lot of value to the business.”

Øyvind joined the keynote at Coalesce 2025 to discuss NBIM's experience with dbt: