The discovery of new catalysts, a process critical to over 80% of globally produced goods, is poised for a revolutionary acceleration. SandboxAQ has announced the release of AQCat25-EV2, a powerful new quantitative AI model trained on a massive dataset of quantum chemistry calculations. This model represents a significant advancement in computational catalyst discovery, enabling researchers to identify new materials for energy, chemical production, and agriculture with unprecedented speed and accuracy.
Quick Intel
SandboxAQ releases AQCat25-EV2, a quantum AI model for catalyst discovery.
It predicts catalyst energetics with near-quantum accuracy but 20,000 times faster.
The model is the first to cover all industrially relevant elements in the periodic table.
It includes crucial magnetic spin polarization data for metals like cobalt and iron.
The technology can accelerate R&D for CO2 reduction, plastic recycling, and fuel cells.
The model is trained on 13.5 million high-fidelity quantum chemistry calculations.
Overcoming a Decades-Long Bottleneck
Traditional laboratory methods for catalyst discovery are notoriously slow, typically processing fewer than 100 candidates per week. This low throughput has limited innovation to incremental changes on existing materials. Quantitative AI models like AQCat25-EV2 break this bottleneck by enabling large-scale virtual screening. It predicts energetics with an accuracy approaching physics-based quantum-mechanical methods but at speeds up to 20,000 times faster, making high-accuracy screening across the entire periodic table practical for the first time.
A Step Change in Model Capability
A key differentiator of AQCat25-EV2 is its inclusion of the quantum effect of spin polarization, a factor crucial for accurately modeling many abundant metals like cobalt, nickel, and iron. This allows the model to expand its accurate range to cover all industrially relevant elements, a first for heterogeneous catalyst models. "AQCat25-EV2 is among the first models that will allow screening in silico on a wide set of chemistries with unprecedented precision and speed," said Dr. Bob Maughon, former CTO at SABIC.
Enabling Broader Discovery and De-risking R&D
The model was developed to provide researchers with the confidence that they are identifying the most promising candidates, not just a narrow subset. Previously, the high cost and time of simulations forced researchers to stop their search after finding a few interesting targets, potentially leaving superior candidates undiscovered. AQCat25-EV2 fundamentally changes this dynamic. "For these industries, we're fundamentally de-risking the R&D process across the entire spectrum of materials science," said Aayush Singh, who leads Catalytic Sciences at SandboxAQ.
The release of AQCat25-EV2 marks a pivotal moment for industries reliant on chemical processes. By providing a tool that combines comprehensive element coverage, critical quantum data, and immense computational speed, SandboxAQ is empowering a new era of materials science where step-change innovations in sustainability and efficiency are within reach.
SandboxAQ is a B2B company delivering solutions at the intersection of AI and quantum techniques. The company's Large Quantitative Models (LQMs) deliver critical advances in life sciences, financial services, navigation, and other sectors. The company emerged from Alphabet Inc. as an independent, growth-backed company funded by leading investors including funds and accounts advised by T. Rowe Price Associates, Inc., IQT, US Innovative Technology Fund, S32, Hillspire Capital, Breyer Capital, Marc Benioff, Thomas Tull, Paladin Capital Group, and others.
TechIntelPro empowers B2B technology decision-makers with the latest industry trends, deep insights, leadership perspectives and actionable intelligence to drive both organizational and professional milestones—and thrive on the tech pulse.