Revolutionising autonomous labs for accelerated materials discovery

Revolutionised autonomous labs for accelerated
materials discovery

Partner with us

Our loop of discovery

Our loop
of discovery

Dunia pioneers a paradigm shift in materials discovery by seamlessly integrating the Design-Make-Test-Analyze (DMTA) cycle. Our innovative closed-loop process marks a departure from traditional methods, leveraging the power of artificial intelligence and robotic precision to redefine the landscape of materials innovation.

Artificial Intelligence

Tailored for real-world data, Dunia’s physics-informed AI integrates principles of physics and empirical verification into its algorithms, providing a robust, insightful approach to problem-solving.

Robotic Automation

Precision meets efficiency as Dunia's cutting edge robotic platform rapidly executes electrochemical experiments, meticulously capturing data to ensure traceable, reproducible results.

Data Engine

Dunia captures terabytes of rich, multi-modal experimental data at scale, transforming fragmented outputs into validated datasets that power predictive models and solve materials discovery’s real-world data gap.

Design

Make

Test

Analyze

Using Dunia’s physics-informed algorithm, our AI accelerates the ideation process, enabling innovative, unbiased material design that can optimise for desired characteristics.

Our robotic tools ensures reproducible and reliable material production. Comprehensive metadata capture guarantees all datapoints can be traced back to the sample of origin.

Our testing set-up is designed for industrial scalability. By simulating relevant processing conditions and timescales, we ensure that all of our results translate seamlessly to industry application.

0101010101010101010101010101

Processing Condition - 1

Processing Condition - 2

Processing Condition - 3

Our testing set-up is designed for industrial scalability. By simulating relevant processing conditions and timescales, we ensure that all of our results translate seamlessly to industry application.

0101010101010101010101010101

Processing Condition - 1

Processing Condition - 2

Processing Condition - 3

Our testing set-up is designed for industrial scalability. By simulating relevant processing conditions and timescales, we ensure that all of our results translate seamlessly to industry application.

0101010101010101010101010101

Processing Condition - 1

Processing Condition - 2

Processing Condition - 3

Our testing set-up is designed for industrial scalability. By simulating relevant processing conditions and timescales, we ensure that all of our results translate seamlessly to industry application.

Our algorithms meticulously dissect and interpret massive datasets, uncovering key insights that drive the next iteration of the design phase; completing the closed-loop cycle and ensuring continuous optimisation.

0101010101010101010101010101

Processing Condition - 1

Processing Condition - 2

Processing Condition - 3

0101010101010101010101010101

Processing Condition - 1

Processing Condition - 2

Processing Condition - 3

0101010101010101010101010101

Processing Condition - 1

Processing Condition - 2

Processing Condition - 3

Optimising for
what matters

Performance

Materials set the boundaries of what technology can achieve. The right material can unlock faster computing, longer-lasting batteries, stronger infrastructure, and lighter transportation. Breakthroughs in materials directly determine the pace of progress.

Stability

Large-scale industrial processes run for years without interruption. Yet, most material candidates are only tested for a few hours. Consequently, lack of stability is one of the most common reason for laboratory results failing to translate to real-world operation. Dunia’s platform focuses on extensive testing and uses lifetime prediction models for early detection of failure modes.

Cost

Scaling innovation depends on materials that are not only high-performing but also practical. Smarter materials reduce waste, streamline manufacturing, and lower overall production costs. By prioritizing efficiency and selectivity, Dunia helps ensure new technologies are economically viable.

Manufacturability

A breakthrough material is only as impactful as its ability to scale. Many discoveries remain trapped in the lab because they rely on rare precursors or complex fabrication. Dunia’s platform integrates manufacturability from the start—tracking abundance, processing requirements, and supply chain considerations to ensure discoveries can move from prototype to production.

Case studies

CASE STUDY
End-to-end discovery campaign
Identifying high-performing catalyst inks within months.
Problem
AEM water electrolysis is an emerging technology that has the potential to unlock the hydrogen economy. Trial-and-error experimentation leads to inefficient experimentation which ultimately hinder commercialization.
Solution
Our autonomous platform used ML-guided design with automated experimentations to quickly identify high performing catalyst ink formulations, outperforming human researchers.
Impact
Our platform outperformed human researchers in increasing energy efficiency while simultaneously minimizing cost. Within just few months, high-performing formulations were identified and validated under industrially relevant conditions.
CASE STUDY
ML-driven Design
Efficient navigation of chemical search spaces.
Problem
Finding the right starting point for a catalyst discovery campaign is difficult as human researchers struggle to screen through the vast amount of scientific data.
Solution
We leveraged unsupervised machine learning techniques in conjunction with domain-specific materials representations to map out chemical search spaces.
Impact
Our data-driven tools enable the efficient navigation of chemical search spaces by exploiting existing knowledge as well as highlighting underexplored areas to search for novel solutions.
CASE STUDY
Automated synthesis
Accessing advanced nanomaterials by tuning processing conditions.
Problem
Particle sizes, distributions and morphology affect the catalytic activity performance of nanostructured materials. Yet, precise control of these properties is difficult to achieve, creating a synthesis bottleneck.
Solution
By tuning the processing conditions of modern synthetic protocols, multi-metallic nanoparticle catalysts were prepared that exhibited high catalytic activity.
Impact
Our platform has access to a large variety of nanostructured materials and can directly map the influence of processing conditions to catalytic function.
CASE STUDY
High-throughput testing
40% increased performance for supercapacitor additives.
Problem
Identifying the right additives is crucial for enhancing the performance and longevity of supercapacitors, but the extensive range of possibilities makes the optimization process difficult and labour-intensive.
Solution
A high-throughput robotic platform was employed to perform electrochemical characterisation and probe the redox behaviour of additives for supercapacitors.
Impact
During an extensive search campaign more than 2500 electrochemical tests were performed, resulting in 45 % increase in capacitance for the champion formulation.
CASE STUDY
Data capture and analysis
100% traceability of research data.
Problem
Machine learning system trained on literature data often yield unsatisfactory performance due to publication biases and missing details in experimental procedures.
Solution
Dunia built a proprietary electrolysis dataset in which meta data are meticulously captured throughout the full sample history. All data is mapped on a domain-specific schema and managed in a cloud database.
Impact
All data are 100% traceable to the sample origin, enabling us to build more powerful machine learning models and immediately validate promising material candidates.

CASE STUDY
End-to-end discovery campaign
Identifying high-performing catalyst inks within months
Problem
AEM water electrolysis is an emerging technology that has the potential to unlock the hydrogen economy. Trial-and-error experimentation leads to inefficient experimentation which ultimately hinder commercialization.
Solution
Our autonomous platform used ML-guided design with automated experimentations to quickly identify high performing catalyst ink formulations, outperforming human researchers.
Impact
Our platform outperformed human researchers in increasing energy efficiency while simultaneously minimizing cost. Within just few months, high-performing formulations were identified and validated under industrially relevant conditions.
CASE STUDY
ML-driven Design
Efficient navigation of chemical search spaces
Problem
Finding the right starting point for a catalyst discovery campaign is difficult as human researchers struggle to screen through the vast amount of scientific data.
Solution
We leveraged unsupervised machine learning techniques in conjunction with domain-specific materials representations to map out chemical search spaces.
Impact
Our data-driven tools enable the efficient navigation of chemical search spaces by exploiting existing knowledge as well as highlighting underexplored areas to search for novel solutions.
CASE STUDY
Automated synthesis
Accessing advanced nanomaterials by tuning processing conditions
Problem
Particle sizes, distributions and morphology affect the catalytic activity performance of nanostructured materials. Yet, precise control of these properties is difficult to achieve, creating a synthesis bottleneck.
Solution
By tuning the processing conditions of modern synthetic protocols, multi-metallic nanoparticle catalysts were prepared that exhibited high catalytic activity.
Impact
Our platform has access to a large variety of nanostructured materials and can directly map the influence of processing conditions to catalytic function.
CASE STUDY
High-throughput testing
40% increased performance for supercapacitor additives
Problem
Identifying the right additives is crucial for enhancing the performance and longevity of supercapacitors, but the extensive range of possibilities makes the optimization process difficult and labour-intensive.
Solution
A high-throughput robotic platform was employed to perform electrochemical characterisation and probe the redox behaviour of additives for supercapacitors.
Impact
During an extensive search campaign more than 2500 electrochemical tests were performed, resulting in 45 % increase in capacitance for the champion formulation.
CASE STUDY
Data capture and analysis
100% traceability of research data
Problem
Machine learning system trained on literature data often yield unsatisfactory performance due to publication biases and missing details in experimental procedures.
Solution
Dunia built a proprietary electrolysis dataset in which meta data are meticulously captured throughout the full sample history. All data is mapped on a domain-specific schema and managed in a cloud database.
Impact
All data are 100% traceable to the sample origin, enabling us to build more powerful machine learning models and immediately validate promising material candidates.

CASE STUDY
End-to-end discovery campaign
Identifying high-performing catalyst inks within months
Problem
AEM water electrolysis is an emerging technology that has the potential to unlock the hydrogen economy. Trial-and-error experimentation leads to inefficient experimentation which ultimately hinder commercialization.
Solution
Our autonomous platform used ML-guided design with automated experimentations to quickly identify high performing catalyst ink formulations, outperforming human researchers.
Impact
Our platform outperformed human researchers in increasing energy efficiency while simultaneously minimizing cost. Within just few months, high-performing formulations were identified and validated under industrially relevant conditions.
CASE STUDY
ML-driven Design
Efficient navigation of chemical search spaces
Problem
Finding the right starting point for a catalyst discovery campaign is difficult as human researchers struggle to screen through the vast amount of scientific data.
Solution
We leveraged unsupervised machine learning techniques in conjunction with domain-specific materials representations to map out chemical search spaces.
Impact
Our data-driven tools enable the efficient navigation of chemical search spaces by exploiting existing knowledge as well as highlighting underexplored areas to search for novel solutions.
CASE STUDY
Automated synthesis
Accessing advanced nanomaterials by tuning processing conditions
Problem
Particle sizes, distributions and morphology affect the catalytic activity performance of nanostructured materials. Yet, precise control of these properties is difficult to achieve, creating a synthesis bottleneck.
Solution
By tuning the processing conditions of modern synthetic protocols, multi-metallic nanoparticle catalysts were prepared that exhibited high catalytic activity.
Impact
Our platform has access to a large variety of nanostructured materials and can directly map the influence of processing conditions to catalytic function.
CASE STUDY
High-throughput testing
40% increased performance for supercapacitor additives
Problem
Identifying the right additives is crucial for enhancing the performance and longevity of supercapacitors, but the extensive range of possibilities makes the optimization process difficult and labour-intensive.
Solution
A high-throughput robotic platform was employed to perform electrochemical characterisation and probe the redox behaviour of additives for supercapacitors.
Impact
During an extensive search campaign more than 2500 electrochemical tests were performed, resulting in 45 % increase in capacitance for the champion formulation.
CASE STUDY
Data capture and analysis
100% traceability of research data
Problem
Machine learning system trained on literature data often yield unsatisfactory performance due to publication biases and missing details in experimental procedures.
Solution
Dunia built a proprietary electrolysis dataset in which meta data are meticulously captured throughout the full sample history. All data is mapped on a domain-specific schema and managed in a cloud database.
Impact
All data are 100% traceable to the sample origin, enabling us to build more powerful machine learning models and immediately validate promising material candidates.

Revolutionising autonomous labs for accelerated materials discovery