QDX opens EXESS quantum chemistry engine to academics

QDX has opened external access to EXESS, its quantum chemistry software engine, and will allow academic researchers to use it at no cost through a project-approval process. Approved academic teams running EXESS on QDX infrastructure will also be eligible for complimentary GPU compute credits.

EXESS, short for Extreme-scale Electronic Structure System, is a GPU-accelerated framework for electronic-structure calculations. QDX says it previously supported a double-precision exascale scientific calculation and set a record for the scale and speed of a quantum chemistry simulation.

Electronic-structure methods sit at the core of quantum chemistry. They calculate how electrons distribute and interact in molecules and materials, helping researchers study bonding, reactivity, and properties that influence macroscopic behaviour.

Computational limits

In drug discovery and materials research, teams often face constraints when modelling large molecular systems with high-accuracy quantum mechanical methods. Many established techniques work best on relatively small molecules, pushing researchers towards approximate models for larger structures such as proteins, polymers, or complex interfaces.

Approximate approaches can struggle to capture key effects in realistic chemical environments, including bond breaking and formation and electronic interactions such as polarisation and dispersion. These effects influence binding, stability, and reaction pathways.

QDX positions EXESS as a way to run more accurate calculations on larger systems than is typically feasible in routine workflows. The engine combines GPU-native algorithms with molecular fragmentation, which divides a large system into smaller components for quantum calculation and then recombines the results to represent the full structure.

According to QDX, this approach extends quantum-level modelling to larger targets such as protein active sites and multi-component molecular interfaces. The company also describes EXESS as suitable for studying chemical reactions and electronic properties in complex systems, including research linked to sustainable energy and materials design.

Record setting

EXESS has been associated with large-scale high-performance computing demonstrations. QDX says the software won the Gordon Bell Prize for what it describes as the largest ab initio quantum mechanical calculation at the MP2 level of accuracy, and that the calculation was more than 1,400 times larger than the previous record.

MP2, short for second-order Møller-Plesset perturbation theory, is a quantum chemistry method that includes electron correlation effects beyond basic mean-field descriptions. It is widely used for its balance of computational cost and accuracy. Scaling MP2 to very large systems has long been difficult because compute and memory requirements grow rapidly as system size increases.

QDX says it is moving EXESS from supercomputing demonstrations into broader research use. The shift also reflects a wider push by specialist computational chemistry firms to expand adoption of GPU-centred simulation beyond national labs and large supercomputing programmes.

Access model

A limited version of EXESS is available immediately through the company's website. Researchers can apply for free academic access through an approval process. Approved projects with clear scientific objectives will receive full access and free GPU time on QDX infrastructure.

This model combines software access with compute allocation, often a decisive factor in whether large quantum chemistry studies can proceed. GPU time can be expensive and difficult to secure for projects that require repeated calculation cycles across many candidate structures or conditions.

QDX is best known for computational drug discovery. The company says it integrates quantum mechanical simulation with supercomputing, artificial intelligence, and drug discovery expertise, and partners with pharmaceutical companies and academic institutions.

Professor Giuseppe Barca, co-founder and head of research at QDX, described the software's design goals and how the engine supports fragmentation approaches for biomolecular work.

"EXESS was built to make high-accuracy quantum chemistry run efficiently on GPUs and to scale across multiple GPUs when time-to-solution matters," said Professor Giuseppe Barca, co-founder and head of research at QDX. "At its core is a highly optimized, GPU-native electronic-structure engine that delivers state-of-the-art performance for conventional quantum chemistry calculations. For biomolecular systems, fragmentation provides the scientifically appropriate pathway to extend this underlying performance to biological length scales, with the GPU engine making those fragment-based calculations computationally practical without compromising accuracy."

Loong Wang, CEO of QDX, said the aim is to increase the volume of quantum chemistry calculations run by the research community by providing both the software and compute for approved academic users.

"We want scientists to have both the software and the compute needed to run as many accelerated quantum chemistry calculations as possible," said Loong Wang. "The more quantum chemistry calculations the world can do, the better. And we will keep working to remove as many barriers as possible."