AI-Driven Chemistry R&D

Atinary’s SDLabs platform significantly accelerates catalysis R&D by using Bayesian Optimization to intelligently navigate vast parameter spaces (including categorical, continuous, and discrete variables), and quickly identifying optimal catalysts, with minimal experimentation. 

In a collaboration with ETH Zurich, the platform enabled the discovery of the best catalyst for CO2 to methanol conversion in just 30 days, representing a 1,000X acceleration compared to traditional methods that would take nearly a century. This was achieved by testing only 0.00072% of all possible combinations, showcasing how AI and data-driven methods lead to faster discovery, maximized conversion and selectivity, and smarter science.

Atinary’s platform supports a wide range of chemical experiments by moving away from traditional trial-and-error methods to an AI-native DMTA-L closed loop. It enables scientists to accelerate and optimize applications like organic synthesis (eg. Buchwald-Hartwig and Suzuki cross-coupling reactions), catalysis and hydroformylation,  formulation, ligand optimization, library screening, and HPLC method development. 

Yes. Atinary maintains strong collaborations with leading companies, including ABB robotics, Agilent, AWS, Bruker, Chemspeed, and Mettler Toledo, among others. Integration with laboratory robotics platforms (such as Chemspeed, Bruker, or IBM RoboRXN), as well as platforms like ACD/Labs’ Katalyst and data management systems like LIMS and ELNs, is optional and seamless via API/SDK. This enables end-to-end closed-loop workflows for Self-Driving Labs® mode, while manual experimentation remains fully supported.

Our SDLabs platform accelerates your chemistry R&D by leveraging agentic workflows and ML algorithms  to automate the various components of planning and running experiments in chemistry, materials, drug development and formulation. We deploy a range of algorithms designed for specific challenges. Some tackle multi-objective problems like balancing yield with sustainability, while others integrate expert knowledge into constrained optimization models. The platform also leverages transfer learning to ensure your past data fuels future success. Supported by autonomous agents that scout literature and diverse datasets, our system is built to minimize overhead and maximize output. This boosts efficiency, productivity, and ROI by moving away from traditional trial-and-error.

When you integrate the SDLabs platform with robotics, you unlock the full potential of a self-driving lab. This automates manual tasks, speeding up processes from hours to minutes and freeing up scientists’ time for critical thinking and breakthroughs, ultimately resulting in a faster time-to-market.

Atinary is committed to state-of-the-art data security and privacy for its SDLabs platform. We are SOC-2 compliant and safeguard all confidential information through AWS hosting and state-of-the-art encryption. Critically, all data and discoveries belong exclusively to the client, ensuring their intellectual property is protected.