™a product by MatterAlchemyTM
AI-guided materials discovery, grounded in physics.
MatterAlchemy's Aether helps R&D teams define a challenge, generate and screen candidate structures, and prioritize what to test next — with every result labeled by evidence level.
Predicted · Computationally validated · Experimentally validated
NVIDIA Inception program member
Material discovery is still slowed by manual, document-driven cycles. Promising candidates can take years to move from hypothesis to a validated material, and failed experiments often take institutional knowledge with them. In the race for electrification, trial-and-error alone is no longer a viable strategy.
Digitizing R&D. From computational prediction to experimental validation.
Aether is a predictive co-pilot that compresses early-stage discovery. It helps teams screen candidate structures, prioritize experiments, and preserve organizational memory — while clearly separating predicted, computationally validated, and experimentally validated results.
Aether is building the digital foundation for the next industrial revolution. We are transforming the physical world of materials into a scalable, generative digital data stream.
From autonomous discovery to enterprise-grade informatics, we're empowering the teams building the future of energy, mobility, and aerospace.
An AI-guided assistant that forecasts properties and thermodynamic stability, helping teams prioritize experiments and reduce avoidable physical iterations.
Leveraging generative models to navigate the combinatorial space of chemistry, with transparent confidence estimates for every candidate.
A single source of truth for your R&D data, preserving institutional memory and eliminating fragmented research silos.
Digitally generate and screen batches of candidate crystal structures for rapid material discovery.
Interactive 3D structural analysis for prioritized candidates, featuring full API connectivity to existing enterprise systems (LIMS/ELN).
A cloud-native infrastructure built to modernize legacy discovery workflows through data standardization and automation.
Harnessing materials informatics to explore stable Li-O structures. Identifying candidate electrolytes and catalysts to address cycle-life degradation in the post-Li-ion era.
Digitizing the search for earth-abundant alternatives. Exploring Fe-N phases to reduce Rare Earth (REE) dependence and geopolitical risk.
Bridging AI and the laboratory. Screening large candidate spaces with stability filters such as Ehull, then ranking structures for experimental validation.
Parameter Setup
AI Inference
Stability Analysis
Final Candidates
Aether connects candidate generation, physics-based screening, and experiment planning in one workflow. Every result carries an evidence status, so teams know what is known and what still needs to be tested.
Model-generated hypothesis. Useful for exploration; not yet tested.
Passed physics-based screening such as stability filters. Not experimentally confirmed.
Confirmed by laboratory work, performed with qualified partners as agreed.
"Aether keeps the evidence trail attached to every candidate, from first prediction to lab result."
| Target Metric | Traditional Approach | Aether MaaS |
|---|---|---|
| Discovery Cycles | Long, sequential experimental loops | Parallel candidate generation and computational screening |
| Physical Iterations | High-cost trial-and-error | Prioritized experiments based on evidence status |
| Data Integrity | Fragmented spreadsheets | Unified Organizational Memory |
| Sourcing Risk | Supply chain REE vulnerability | AI-guided substitution exploration |
| Digital Foundry | Isolated R&D Silos | Shared, searchable discovery memory |
The Digital Foundation for Discovery to Delivery
Materials Informatics SaaS
Self-serve access via multi-tenant subdomains. Ingest legacy data and run autonomous structural inferences on Azure ML.
Internal R&D / IP Licensing
We develop internal discovery programs (e.g. Li-O/Fe-N). Clients can license computationally evaluated structural candidates.
Joint Development Partner
Deep collaboration using custom adapters fine-tuned on proprietary client (often failed) experimental data.
From challenge definition to experimental validation
Evidence status
Model-generated hypothesis, not yet tested.
Passed physics-based screening; not experimentally confirmed.
Confirmed by laboratory work, when that work is in scope.
Set target properties, elements, and constraints.
InputGenerate candidate crystal structures from the defined search space.
PredictedApply physics-based filters such as Ehull.
Computationally validatedReview stability and feasibility; lab validation where available.
Lab-dependentRank candidates by stability, synthesizability, and constraints.
Computationally validatedFeed results back into the model and refine the next round.
FeedbackComputational results are provided to prioritize experiments. They must be experimentally validated before industrial use.
De-Risking the Supply Chain through Digital Innovation
As global regulations tighten on rare earth mining and "conflict minerals," the ability to design alternatives digitally is a critical competitive advantage.
Aether enables organizations to automate the search for earth-abundant alternatives, providing quantifiable ESG metrics and supply chain predictability from the earliest stages of R&D.
Discover Fe-N alternatives to high-HHI magnets
Transition to domestic, earth-abundant sourcing
Quantifiable ESG metrics for stakeholders
Stable pricing, no commodity volatility
Presenting our work within the scientific and research community
7th International Conference on Molecular Simulation · IISc Bengaluru
Girish M
Presented a talk/poster at the 7th International Conference on Molecular Simulation, held at the Indian Institute of Science, Bengaluru.
7th International Conference on Molecular Simulation · IISc Bengaluru
Hemanth Shetty Mundaje
Presented a talk/poster at the 7th International Conference on Molecular Simulation, held at the Indian Institute of Science, Bengaluru.
Built by domain experts with deep roots in AI infrastructure and enterprise software
CEO & Co-founder
Brings deep expertise in building scalable AI/ML platforms, previously at Microsoft driving enterprise-grade software solutions. Holds an MS in Computer Science from Northeastern University with prior experience at VMware, New York Life Insurance Labs, and Tesco. Leads product vision, fundraising, and go-to-market strategy.
LinkedIn
CTO & Co-founder
Over a decade of software engineering experience across enterprise, government, and open-source ecosystems. Built OpenProject integrations, developed microservices at Cisco (WebEx Teams), contributed to the TMT observatory project at IIA, and TA at IISc. Architects the Aether platform and leads engineering delivery.
LinkedInTell us what you are trying to discover. Share the material class, constraints, and current baseline, and we will respond with a focused next step.
Please do not include confidential information in this form. We can put an NDA in place before a detailed technical exchange.