Seismic data contains valuable information about the subsurface, but the datasets are often too large and complex to be fully utilised. At NR, we have developed an AI-based foundation model that enables more efficient workflows, more consistent analyses and better use of subsurface data. contribute to more efficient work processes, consistent analyses and improved exploitation of subsurface data.<\/strong><\/p>\n\n\n\n Artificial intelligences is already used for a wide range of tasks in seismic interpretation and geoscience, from the automatic detection of faults and other geological structures to data segmentation and classification. Such models can be highly effective, but they are typically developed for a single specific task and require large amounts of labelled training data. <\/p>\n\n\n\n Foundation models offer a new approach. Rather than being trained for a specific task, the model is first trained on large amounts of raw data to learn general patterns and relationships within the data. It can then be adapted to wide range of applications with significantly less additional training and far fewer labelled data. <\/p>\n\n\n\n Traditional AI models can be compared to specialist tools: they solve a particular problem well, but often need to be rebuilt if the problem changes.<\/p>\n\n\n\n A foundation models function more like a platform. Once it has learnt to understand structures and patterns, it can serve as a starting point for many different applications. This reduces development time, lowers costs and makes it possible to explore new methods and application areas more quickly than before. <\/p>\n\n\n\n Because the most resource-intensive work has already been carried out in the foundation model, smaller and more specialised models can be built on top of it for specific tasks. <\/p>\n\n\n\n Foundation models can help to:<\/p>\n\n\n\n The NCS model is a foundation model trained on seismic data from DISKOS, Norway’s national data repository, using the Norwegian supercomputer Olivia. The model has been developed in collaboration with Aker BP and Equinor and is available as open-source software. <\/p>\n\n\n\n Although the model has been developed using data from the Norwegian Continental Shelf, the technology can be applied in far beyond oil and gas exploration. Potential applications include carbon storage, offshore wind and other offshore developments where knowledge of the subsurface is essential. <\/p>\n\n\n\n By making the model available as open source, we enable both research organisations, companies and other users to adopt the technology, further develop it and explore new areas of application. <\/p>\n\n\n\n Whether you are looking to adopt existing technology, explore new applications or develop solutions tailored to your own challenges, we would be happy to discuss potential opportunities for collaboration.<\/p>\n\n\n\n![\"Ocean](\"https:\/\/nr.no\/content\/uploads\/sites\/2\/2026\/06\/grunnmodell-bransje-1024x683.jpg\")
What are foundation models?<\/h2>\n\n\n\n
From specialist tools to advanced platform technology<\/strong><\/h2>\n\n\n\n
Benefits of foundation models<\/strong><\/h3>\n\n\n\n
\n
The NCS model \u2013 a foundation model trained on Norwegian data<\/h2>\n\n\n\n
Applications and collabortion<\/strong><\/h3>\n\n\n\n
![\"\"](\"https:\/\/nr.no\/content\/uploads\/sites\/2\/2024\/05\/anders-ueland-waldeland-10.jpg\")
\n\t\t\t<\/figure>\n\t\t\t\t