Insilico Medicine and Saudi Aramco Launch ‘sorbaMOF DB’ to Accelerate AI-Driven MOF Discovery, Advancing CO₂ Capture

August 17, 2026

Insilico Medicine Hong Kong

Overview

Insilico Medicine and Saudi Aramco have launched ‘sorbaMOF DB,’ advancing AI-driven MOF (metal-organic framework) discovery. This validated, extensible materials database and benchmarked computational protocol ensure reliable CO₂ adsorption predictions. The initiative aims to establish a high-quality data foundation for generative AI and foundation models, accelerating sustainable technology development like CO₂ direct air capture (DAC) technologies.

In Depth

Key Findings

Insilico Medicine and Saudi Aramco have announced ‘sorbaMOF DB,’ an innovative dataset and computational protocol designed to dramatically accelerate AI-driven MOF (metal-organic framework) discovery. This initiative aims to establish a validated and extensible materials database that ensures reliable CO₂ adsorption predictions, thereby building a high-quality data foundation for generative AI and foundation models.

Technical / Clinical Details

sorbaMOF DB is more than just a materials database; it is a comprehensive platform that supports the entire lifecycle of AI-driven MOF discovery. Its key features include:

Validated and Extensible Database: Constructed through a rigorous validation process, it contains abundant high-quality data on MOF composition, structure, and CO₂ adsorption properties. It is designed to be extensible with new discoveries and experimental data.
Benchmarked Computational Protocol: Computational methods for predicting MOF CO₂ adsorption capabilities are evaluated against multiple criteria, ensuring their reliability. This enhances the learning efficiency and prediction accuracy of AI models.
Data Foundation for Generative AI and Foundation Models: This database provides high-quality data for training generative AI models (e.g., AI that autonomously generates new structures) for MOF design and discovery, as well as foundation models applicable to a wide range of materials science tasks.
Reliability of CO₂ Adsorption Prediction: The accuracy and reliability of AI’s predictions for MOF CO₂ adsorption properties are significantly improved. This enables rapid identification of MOFs that can efficiently capture CO₂ under specific conditions (e.g., specific temperatures and pressures).

This technology empowers AI to efficiently explore the diverse structures, often called ‘molecular Legos,’ of MOFs, enabling the improvement of existing MOFs or the discovery of entirely new ones. CO₂ Direct Air Capture (DAC) technology, in particular, is a key technology for climate change mitigation, and sorbaMOF DB significantly boosts progress in this field.

Background & Context

MOFs, as porous materials, hold great potential across a wide range of applications, including gas separation, storage, catalysis, and sensing. CO₂ capture and storage, in particular, is gaining global attention as a climate change mitigation strategy, and MOFs are highly promising due to their high adsorption capacity. However, efficiently discovering the optimal MOF for specific applications (e.g., DAC) from millions of theoretical candidates has been challenging with traditional scientific methods. The merger of Insilico Medicine’s expertise in AI-driven drug discovery and Saudi Aramco’s commitment to R&D is bringing new solutions to this field.

Strategic Significance & Outlook

The announcement of sorbaMOF DB signifies the full-scale deployment of AI-driven materials discovery in the MOF field. This high-quality data foundation and computational protocol will contribute to reducing costs and improving the efficiency of CO₂ Direct Air Capture technologies, accelerating climate change mitigation. Furthermore, this approach is expected to be applied to other MOF applications (e.g., hydrogen storage, drug delivery), further enhancing the importance of AI and data in overall materials science research. In the future, it holds the potential to contribute to the realization of a ‘closed-loop’ materials discovery ecosystem, automating everything from MOF design to synthesis, in collaboration with autonomous lab systems.

Source: https://insilico.com/news/5kip5kk0s1-from-molecular-lego-to-high-quality-data

Get our weekly technology intelligence — free

Receive an infographic that lets you judge at a glance whether each field’s analysis report is worth reading.

Subscribe Free — Weekly Tech Intelligence

By subscribing, you’ll receive Troy-Technical’s weekly technology intelligence newsletter.