The fusion of Artificial Intelligence (AI) and materials science is changing how industries work, from energy and construction to advanced manufacturing. AI is speeding up discovery, from low-carbon cement to quantum-designed materials. To keep up, your IP strategy needs to evolve, too.
As companies harness AI to accelerate material discovery and optimise production processes, securing robust patent protection becomes critical to safeguard these technological breakthroughs.
Protecting innovations in materials science is often no longer as simple as focusing on the purely chemistry aspects. It's about algorithms, data, and design processes, too.
Real-World Examples
The Materials and Manufacturing Showcase recently highlighted the impact of a £1.75 billion investment in materials and manufacturing. It spotlighted pioneers demonstrating how AI transforms material development. Below, we analyse three standout innovators shaping this landscape and the patent strategies essential for safeguarding AI-driven advancements:
1. Matnex (Materials Nexus): Quantum-Powered Material Discovery
Matnex's 2025 partnership with Viridien to build the world's largest AI supercomputer for materials science marks a paradigm shift in R&D efficiency. Their platform enables:
- AI-driven materials development for accelerating the discovery and design of new materials.
- Decades-to-months R&D compression through AI-driven design of sustainable materials for energy storage, transport, and computing.
- Sustainable material design with tailored electronic/chemical properties.
2. Material Evolution: AI-Optimised Low-Carbon Cement
This innovator's AI-driven cement formulations address the cement industry's 8% global CO₂ emissions through:
- Specialising in ultra-low-carbon cement, ensuring a sustainable supply chain.
- Feedstock optimisation, enhancing material durability while reducing carbon footprint.
- Advancements in computational materials design and AI's role in advancing sustainable materials.
3. HAL Robotics: AI-Driven Manufacturing Precision
While primarily a robotics specialist, HAL collaborates on materials science initiatives through partnerships. It demonstrates IP opportunities at the AI-materials interface:
- AutoCorrect is an autonomous system for defect detection and correction in manufacturing. This tool uses computer vision, deep learning, and AI to analyse surface defects (e.g., cracks, burrs) and generate robotic repair processes without human intervention.
- Partnership with institutions like the Build'In Co-Innovation Lab at École des Ponts ParisTech, which researches new materials and construction methods, providing robotics training and software support for these material science initiatives.
- Their work with 3D concrete printing and timber machining in research projects further demonstrates indirect involvement in materials science through applied robotics.
This landscape underscores how AI is redefining materials science, where patents can be used to protect algorithmic breakthroughs alongside novel compounds. AI-driven material innovations can qualify for patent protection in Europe, provided they extend beyond merely applying established AI algorithms to automate or enhance conventional processes.
Strategic Patent Drafting for AI-Material Innovations
When protecting AI-driven material inventions in Europe, we prioritise three critical considerations:
1. Technical Character & Inventive Step
A core requirement is to ensure that the invention demonstrates a technical effect beyond abstract AI algorithms or mathematical methods:
- Focus claims on specific technical applications (e.g., designing energy storage materials with optimised electronic properties).
- Use the problem-solution approach when drafting and prosecuting a patent application to emphasise the technical contribution of the invention, such as how the AI solves material-specific challenges.
- Include experimental data and/or simulations to demonstrate unexpected improvements over conventional methods.
2. Sufficiency of Disclosure
Patent applications lacking sufficient detail on AI training risk refusal under Article 83 EPC. It is important to:
- Detail AI architecture (model type, layers), training data (sources, size, outlier handling), and validation protocols.
- Include details of implementation to demonstrate practical utility.
- Specify hyperparameters, activation functions, and optimisation techniques to ensure reproducibility.
3. Claim Drafting
Strategic drafting of claims helps to manage the scope of protection by:
- Filing initial claims covering AI methods broadly and independently of specific use-cases, then narrowing during prosecution.
- Drafting hybrid claims where appropriate, combining material composition claims with AI architecture details.
- Include fallback positions in dependent claims tied to specific applications (e.g., battery materials).
Protecting Innovation at the AI-Materials Frontier
By addressing these factors, applicants can navigate the EPO's stringent requirements and secure robust protection for AI-driven materials innovations. Keltie's dual expertise in computational materials science and the EPO AI guidelines positions clients to:
- Protect novel AI architectures applied to material-specific challenges
- Secure manufacturing implementations through process claims
- Future-proof portfolios against evolving examination practices
The convergence of AI and materials science is moving fast, and so is the patent landscape. For innovators working at this cutting edge, strategic IP advice isn't optional; it's essential.
Whether you're training AI to discover compounds or deploying it in industrial applications, we're here to help you protect what makes your innovation unique.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
