Join Molecular Universe MU-1 Live Demo

Molecular Universe first became known for its molecular databases. Having the world’s largest databases of battery relevant molecular properties at one’s fingertips completely revolutionized battery material discovery. In Molecular Universe (MU-1), we are expanding the databases to cover most liquid and solid battery electrolytes, including solvents, additives, diluents, salts, and solid state materials. We are offering 1M database fully viewable to the Public, and fully searchable to Research, Explorer and Team tiers, and 200M database fully searchable to Enterprise and Joint Development tiers with more in-depth properties.

As one of the most popular tools in Molecular Universe, Ask is like having world-renowned battery scientists and engineers at your fingertips. MU-1 enhances Ask performance by integrating the latest GPT-5 models and SES proprietary training data including publications, patents, and human intuition. We are offering Lightning to the Public, and more advanced Pro and multi-agent Deep Space to Research, Explorer, Team, Enterprise and Joint Development tiers. Deep Space now has senior scientist-level agentic capability to understand users’ queries and recommend full solutions including detailed reasoning, experiments and predictions. Deep Space is also fully integrated with Maps, Predict and other features in MU-1, and it is one step closer to full agentic capability that can outperform the very best human battery scientists.

As another popular tool in Molecular Universe, “Find-Friends” helps users find other molecules with similar physicochemical and/or structural properties as the input molecule, which is invaluable in helping users quickly broaden their horizon of possibilities in establishing new IPs. In MU-1, we are offering “Intelligent Find-Friends” that also takes into account the battery chemistry environment and desired performance for the molecules we are seeking, quickly identifying the most relevant “friends”.

Having mapped the world’s largest molecular databases, for the first time, we are venturing beyond the molecular universe and into the formulation universe, which involves much higher dimensions of complexity. Electrolytes are more than just single molecules; they are formulations consisting of ions dissociated and solved by solvent molecules, which interact in various manners among themselves. Here we have developed advanced computation chemistry tools such as molecular dynamics simulations driven by polarizable force fields. These tools have achieved unprecedented accuracy, scale and speed for prediction of formulation level properties including viscosity, solubility, miscibility, conductivity, and more. Users can enter their desired formulations, and we will compute their properties.

Having explored the formulation universe, we continue our journey into the cell universe, an even higher dimension of complexity. Ultimately, cell level performance is what users care about most, since the final product is neither a molecule nor formulation, but a complete cell. We are truly in uncharted territory, as currently there are no known methodologies for connecting molecule and formulation properties to cell performance; the interfaces and interphases are simply beyond what first-principle can predict. Fortunately, the power of AI and machine learning, grounded in our massive effort to meticulously collect cell experimental performance data over diverse chemistries and as functions of different molecules and formulations, is showing promise. Users can select a particular cell chemistry and different unknown molecules from the Molecular Universe to see their impact on cell performance and accurately predict cycle life. Users can even blindly input data of early cell cycle life, and MU-1 will predict end of life, all without any prior knowledge of the input cell chemistry or test environment.