The next step for National Data Repositories

DISKOS recently turned 30, and it got me thinking about the role of Na­tional Data Repositories (NDRs) today, and how they can stay relevant as times change.

Historically, NDRs served as a custodian of posterity – a place where seismic records and well logs went to be archived, protected by regulatory mandates but often isolated from the pulse of active operations. However, as the industry grapples with the accel­erating complexity of the energy tran­sition, this archival model is proving insufficient.

Currently, NDRs are more impor­tant than ever as data continues to be a fundamental bottleneck for the growing AI industry. Every digital twin, simulation, and predictive mod­el relies on a pipeline of standardised, accurate data; if that foundation is fractured, the entire analytical super­structure collapses.

The Norwegian Offshore Directo­rate has always been very forward-lean­ing when it comes to technology, and as such, it comes as no surprise that over 30 years, they have taken the NDRs from a simple national data repository for exploration and production-relat­ed data to a platform complete with AI-powered chatbots. Strategic invest­ments in data capture, such as the Re­leased Wells Initiative, AVATARA-P and Bypassed Pay studies, have enriched the existing data with modern insights, new datasets and strong foundations for AI and machine learning projects.

Operators and service companies have used each of the aforementioned projects to better understand the Nor­wegian Continental Shelf, produce new breakthroughs in interpretation such as automated biostratigraphy, seismic foundation models and more. One of the most interesting things to come out of the digitisation and standardisation process has been the relevance and ease of use of this infor­mation to support energy transition projects such as carbon storage.

To me, the next natural step on this journey would be for the NDR to curate open models, benchmarks and other important architectures for supporting AI projects. By hav­ing nationally realised and published benchmarks and guidelines, I believe it would be easier to build trust in the results and also accept generative in­put for decision support, automation and more. Perhaps there’s even a place for NDR to provide the means to gen­erate realistic synthetic data.

For regulators and national data managers, this means we could move away from intense pre-qualifications and licence rounds and toward smart recommender systems that match the right asset to the right partner in real time. Imagine a world where agentic AI performs the matchmaking, ana­lysing a company’s technical DNA and financial appetite against the specific geological risks of a block to suggest the most viable partnerships before a single bid is even cast.

Where does this end?

I think the logical endpoint is a radical simplification of the subsur­face interface. If the NDRs solve the problem of data acting as a bottleneck, then we may soon see a world where the NDR merges with frameworks like OSDU to become a universal backend. From here, organisations and individuals could connect self-built applications, agents or even build toolkits on the fly – perhaps bypassing interpretation software entirely.