Palo Alto, CA, June 12, 2025 (GLOBE NEWSWIRE) -- Aionics, Inc., a leading developer of AI-powered materials discovery technologies, today announced the launch of its Artificial Molecular Intelligence (AMI)platform—a first-of-its-kind capability for real-time, autonomous formulation screening. This breakthrough marks a new era in chemical R&D, offering transformative speed and scale in the design of high-performance battery electrolytes and a wide range of other chemical formulations.
With AMI, Aionics introduces a new class of artificial intelligence purpose-built for the staggering combinatorial complexity of chemical mixtures and their integration into real-world devices. Unlike conventional AI tools that analyze molecules in isolation, AMI evaluates full formulations—including molecular interactions, interface behaviors, and commercial constraints—in milliseconds. The platform currently screens over 3,000 formulations per second, enabling rapid optimization across a library of more than 10 billion molecules, including both commercially available compounds and novel molecules created through generative AI.
“Ami is a transformative shift in materials development. We’re no longer bound by trial-and-error or narrow intuition-led approaches,” said Austin Sendek, CEO of Aionics. “We’ve created a system that can autonomously explore chemical space and design formulations optimized for performance and practicality—at a pace the industry has never seen before. Imagine being able to screen any and all molecules for any and all applications. This represents a profound, exciting shift for the chemicals and materials sector.”
Introducing the Ami Platform
To commercialize AMI, Aionics has built the Ami platform (pronounced “Amy”), an integrated platform combining high-speed simulation, machine learning, and lab feedback in a closed-loop architecture. The Ami platform allows customers to specify performance targets and then rapidly returns optimized formulations that meet both technical and commercial criteria.
Ami addresses the two core challenges holding back formulation design today: reliable computational models for predicting key properties of interest do not exist for many key application areas, and the breadth of chemical space is simply too large to ever explore without AI acceleration. This is especially true in the space of battery electrolytes, where most commercialized Li-ion electrolytes are made up of different combinations of the same 11 molecules, while there over 10 billion unexplored molecules.
The Ami platform is already designing customized, high-performance electrolyte formulations today, with expansion into new markets on the horizon.
Key features include:
- Massive molecular library – Over 10 billion molecules, including all commercially procurable compounds and new molecules generated via proprietary generative models, building on intellectual property from the ARPA-E DIFFERENTIATE program.
- Closed-loop optimization – Ami’s predictions are continuously improved through experimental validation using Aionics’ in-house testbeds.
- Commercial filtering – Formulations are screened not only for performance but also feasibility, using models that predict price, synthetic accessibility, vendor availability, and more.
This tightly integrated approach ensures that only practical, scalable, high-performing formulations are forwarded for laboratory validation.
Built for Speed and Scale
Ami’s core technology is the result of more than a decade of research by Chief Executive Officer Austin Sendek and Chief Scientist Venkat Viswanathan, whose academic work on AI for materials design has been supported by millions of dollars in grant funding from agencies including DOE ARPA-E. Together, they have published over 150 peer-reviewed papers and hold dozens of patents in the fields of energy storage, machine learning, and quantum chemistry.
Key innovations include:
- DiffMix, a differentiable machine learning framework combining physics-informed corrections with graph neural networks to predict non-linear mixture behavior.
- Surrogate modeling for quantum simulations, accelerating traditional DFT by up to 1 million-fold for interfacial predictions such as battery cycle life.
- A hybrid simulation strategy using Molecular DFT, MLIPs, MD, and proprietary quantum descriptors to generate high-information features from low-data environments.
These capabilities are supported by a targeted compute infrastructure designed to balance scalability and precision. Aionics performs intensive feature generation, model training, and inference through a suite of GPU-optimized workflows. The further development of these workflows is supported by Aionics’ membership in the NVIDIA Inception Program and Google Cloud for Startups.
Real-World Impact
Within the battery electrolyte space, Ami supports customized electrolyte development for a range of customer needs, including long cycle life batteries for space applications, high- and low-temperature-stable cells for automotive, high-power cells for electric aviation, and nonflammable cells for any application where safety is critical.
“Electrolytes control six of the eight critical properties of a battery, ” said Viswanathan. “With Ami, our partners can unlock performance gains and safety improvements quickly—and without changing the rest of their cell design. As one example, we recently found multiple strategies to enable a customer to increase the safety and power of their electrolyte simultaneously, which was previously thought of as a zero-sum tradeoff.”
While most of Aionics’s customer work remains confidential, the company has previously announced its commercial partnership with Cellforce Group, the electric vehicle battery subsidiary of Porsche. In addition, Aionics recently announced a partnership with an aerospace company to develop a new aviation grade nonflammable electrolyte that increases battery stability and adheres to stringent aviation safety standards.
“Ami is not just helping us to screen molecular libraries more quickly – it is enabling us to break past traditional performance boundaries in electrolyte design,” Viswanathan added.
Leadership in LLM-Powered Chemistry
In 2023, Aionics became the first company to publicly announce a battery science-specific large language model, which received significant market coverage at the time. Since then, the company has expanded this effort through securing licensed access to tens of thousands of peer-reviewed scientific papers on battery design from a major publisher.
Aionics’ LLM stack serves as a scientific co-pilot within Ami, enabling automated literature analysis, experiment design assistance, and complex workflow automation.
Looking Ahead
Aionics plans to expand the AMI platform to support other formulation-intensive industries, including rocket fuels, industrial and nuclear coolants, lubricants, specialty chemicals, and beyond. In parallel, the company is completing the internal rebuild of a self-driving robotic test-stand originally developed in Viswanathan’s University of Michigan lab, enabling human-out-of-the-loop experimental validation of formulation candidates.
About Aionics, Inc.
Founded in 2020, Aionics is pioneering a new era of formulation design by combining artificial intelligence, quantum simulation, and proprietary data to develop custom, high-performance materials for mission-critical applications. Headquartered in Palo Alto, Aionics works with leading companies in automotive, aerospace, defense, and energy to design drop-in chemical solutions that improve performance, safety, and sustainability.
To learn more, visit www.aionics.io
Aionics, Inc. aionics@v2comms.com
