featured image

News

Shearwater unlocks AI-accelerated risk insights for seismic data with NVIDIA technology

21 Aug 2025

polygon iconWe’re bringing next-gen clarity and performance to the subsurface. By combining our seismic and AI expertise with NVIDIA’s accelerated computing platform, we’re changing how uncertainty is handled so energy decisions can be made with greater confidence and reduced risks.

We’re announcing a significant milestone in our multi-year collaboration with NVIDIA, delivering breakthrough progress in applying Generative AI (GenAI) to Uncertainty Quantification (UQ) for seismic processing and imaging.

This paves the way for our processing software Reveal where users can access advanced NVIDIA-accelerated tools for quantifying seismic-related uncertainty and strengthening risk assessment in exploration, production and carbon storage projects.

By harnessing NVIDIA’s accelerated computing platform, the two companies are demonstrating the potential to leverage industry-first, fit-for-purpose GenAI architecture to efficiently learn the uncertainty of any given seismic data challenge and generate user-chosen outputs, for instance different subsurface model scenarios.

The new GenAI-based UQ approach is designed to integrate with Shearwater’s processing and imaging software Reveal─accelerated by the NVIDIA GH200 Grace Hopper Superchip and NVIDIA CUDA-X libraries─bringing power-efficient, user-ready risk assessment capabilities to clients across energy and new energy markets.

“Our collaboration with NVIDIA continues to unlock groundbreaking capabilities for geophysicists across the energy industry,”
says Simon Telfer, SVP Technology, Software, Processing & Imaging of Shearwater. “By combining our seismic expertise with NVIDIA’s full-stack AI platform, we are creating tools that not only advance geoscience but also empower our clients to make more confident decisions in a rapidly changing energy landscape.”

AI seismic risk insights . Two side-by-side seismic acoustic impedance models labeled Scenario A and Scenario B. Both are subsurface models generated from the same data using GenAI-based uncertainty, showing variations in geological layers represented in orange, brown, and blue tones

This development reinforces Shearwater’s pioneering role in applying advanced AI and high-performance computing to seismic science, paving the way for safer, more efficient, and more responsible energy and carbon storage projects.

Building on significant performance gains with NVIDIA and the ongoing, close-knit GenAI architecture development between our teams, this breakthrough marks the next chapter in the Shearwater and NVIDIA collaboration.