First Published: AAPG Explorer - December 2025, by Meher Gajula, Keyla Gonzalez, Ben Lasscock, Sathiya Namasivayam, Altay Sansal and Alejandro Valenciano, TGS
Introduction
Oil and gas exploration grapples with an awkward equation: the data grows, but the day does not. Cloud platforms have changed availability, with petabytes of SEGY, LAS files, and reports stored in systems that virtually never go down. AI now changes the rate and scale of what we can learn from this data. With columnar, cloud-native formats like open source, MDIO and Parquet, our scientific data is not only stored, but also computable and teachable.
Automated salt interpretation trained in the Gulf of Mexico (America) and Brazil and applied in Angola. The model shows state of the art generalization “out of domain.”
The new inflection is the combination of generative AI and domain-specific foundation models, working together with a new generation of automated AI agents. These systems don’t merely search; they connect, linking survey seismic data with other metadata, processing histories, entitlements, and interpretations. Creating business content becomes shorter, clearer, and, crucially, repeatable at an enterprise scale.
Read the full article here.