Datarock has announced the beta launch of Datarock Chip, a first of its kind machine learning system that extracts structured geological data from RC and Aircore chip imagery.
The technology addresses a persistent challenge in exploration: turning subjective, manual chip logging into quantitative, auditable datasets that can be analysed at scale.
For decades, chip logging has been plagued by inconsistency. Results vary depending on who’s doing the logging, lighting conditions on site, even time of day. That variability makes it hard to compare data across campaigns or use historical chip imagery for anything beyond visual reference. Datarock Chip changes that by applying computer vision and machine learning to standardise the process, identifying, segmenting, and analysing chip photography to extract metrics on colour, morphology, and texture with precision.
The outcome is simple: analytics-ready datasets in hours instead of weeks. Data that’s consistent across geologists, comparable between drilling campaigns, and ready to integrate with Datarock Core or other geological systems.
“For our customers, time and confidence in data quality are everything,” said Thomas Picherit, Chief Commercial Officer at Datarock.
“Datarock Chip bridges the gap between imagery and measurable geology, giving exploration and production teams earlier visibility of geological trends and improving the ROI on every metre drilled. It builds on the Datarock Core foundation to make data-driven decision-making even faster and more accessible
across operations.”
Dr Rian Dutch, Head of OBK Solutions at Datarock, added, “Chip datasets have always contained huge untapped potential beyond just a geochemical sample. Given that approximately thirty million metres of RC drilling is conducted by companies each year, it is remarkable how little structured insight is typically extracted from that data. Datarock Chip changes that by using machine learning to quantify what geologists have always observed qualitatively. It is a genuine game changer for how we interpret, use and integrate RC data.”
The system outputs interval-based CSVs and downhole plots that visualise colour distribution, grain characteristics, and texture clusters. Those datasets plug directly into domaining workflows, geotechnical analysis, or machine learning pipelines. Chip data can also be fused with geochemistry, hyperspectral scans, and MWD logs to build multivariate models for properties like lithology prediction, hardness estimation, or mineralisation targeting.
Datarock Chip will launch in beta with a single subscription tier built for flexibility, from single-rig pilots to enterprise-scale rollouts. It includes access to continuously updated models, secure cloud storage, and integration with Datarock’s NEXUS infrastructure.
Datarock Core customers will gain early access to Datarock Chip, giving them a first-user advantage during the beta phase.
The team is extending Datarock Chip capability beyond chip trays to capture value from other storage and sampling formats used in exploration. If clients have a specific use case or dataset, Datarock welcomes the opportunity to analyse it as part of its ongoing development and collaboration.
For more information, visit Datarock’s website.
