Engineers have long understood the properties they need from materials, such as magnetic performance, conductivity and thermal resistance. But translating those requirements into real, manufacturable materials has traditionally been hard to achieve. Here, Gustavo Regueira Llansó, director of strategy and operations at Altrove AI, one of CWIEME Berlin’s Innovation Zone exhibitors, explains why the materials science community has been experiencing the same problem for decades. 

We know what “good” looks like. Engineers understand the properties they need, whether it’s magnetic properties, conductivity or thermal resistance. But translating those requirements into real, manufacturable materials has often been slow and expensive. However, that is now changing as AI completely changes how we think about materials.

This change could not come at a more critical moment. Across electrical engineering, demand for key raw materials has surged. Copper, rare earths, graphite and specialised alloys are increasing in demand significantly. The electrification of transport, the expansion of renewable energy and the digitalisation of industry are all driving unprecedented consumption. At the same time, supply chains have become more fragile and more concentrated, exposing manufacturers to risks that were once easy to overlook.

Not long ago, materials could be treated as a procurement issue. Today, they sit at the centre of business strategy.

Manufacturers are increasingly aware that their ability to produce depends on forces far beyond their control. Finding new suppliers or diversifying sourcing regions are still necessary, but they are no longer sufficient. In many cases, the material itself has become the problem.

This is where AI is having a huge impact. Historically, industries have relied on two main levers to address material challenges: extracting more resources or recycling what already exists. AI is now introducing a third path – possibly the most transformative one – creating entirely new materials tailored to specific needs. Now, instead of asking, “Where can we source this material?”, manufacturers can ask, “Can we design a better one?”

AI enables this by dramatically expanding the range of possibilities. Where scientists once worked with a relatively small set of known material structures, machine learning models can now explore millions of potential configurations. More importantly, they can predict how those materials will behave before they are even synthesized. Properties that once required months of simulation can now be estimated in a fraction of the time.

Knowing the properties of a material is only part of the challenge. Like identifying the ingredients of a recipe, the real difficulty lies in defining the precise process needed to produce it at scale. In materials science, this “recipe” is the synthesis process: the precise steps required to turn a theoretical material into something that can be manufactured at scale.

By combining predictive models with experimental feedback, it is now possible to generate and refine production processes far more efficiently. Instead of relying on slow, manual iteration, companies can use AI to propose multiple “recipes,” test them rapidly and learn from the results. Even failed attempts become valuable data, feeding back into the system and improving future outcomes.

AI is also changing how innovation itself happens. In the past, new materials were often developed in isolation within research labs or academic institutions and only later matched to industrial use cases. This “push” model frequently led to misalignment between what was created and what the market actually needed.

Today, manufacturers can define their requirements upfront. They can specify the exact properties needed, whether these are driven by performance targets, cost constraints or supply chain considerations. AI systems can then work backwards, identifying and      designing materials that meet those criteria.

This “pull” model represents a significant shift in how materials are developed. While the concept itself is not new, AI makes it far more practical by enabling manufacturers to define precise requirements and rapidly explore viable material solutions that meet them.

It also challenges one of the industry’s most persistent assumptions: that trade-offs are inevitable.

For years, material selection has been a balancing act between cost, performance and sustainability. Improving one often meant compromising another. But by exploring a vastly larger design space and optimising across multiple variables simultaneously, AI is beginning to break this assumption. It is increasingly possible to develop materials that are high-performing, cost-effective and more sustainable.

This shift has direct implications for supply chains. Volatility in raw material markets is a huge issue for manufacturers. Price fluctuations, geopolitical tensions and supply disruptions can all have immediate operational impacts. By enabling the development of alternative materials, AI applied to materials science introduces flexibility into a system that has long been rigid.

When companies have more options, they are less dependent on any single source. Rather than simply managing dependencies, manufacturers now have the opportunity to move beyond them altogether. By developing alternative materials tailored to their needs, they can reduce reliance on constrained resources and fundamentally change their supply strategies.

Companies that embrace AI-driven material innovation early have the chance to turn a problem into something that gives them a competitive advantage. Developing or adopting novel materials can not only solve immediate supply issues but also help differentiate products in the market. In some cases, it can even redefine entire categories of performance.

As more organisations begin to explore this space, the pace of discovery will accelerate, creating a virtuous cycle of innovation across the industry. Also, with new commercial models emerging, companies can explore AI-driven material innovation with minimal upfront investment, while benefiting from significant upside potential.

This is why the conversation around raw materials is becoming central to events like CWIEME Berlin. What was once a niche technical topic is now a priority for electrical engineering as a whole. The future of the industry will be shaped as much by materials as by design or manufacturing processes.

At Altrove, we see this shift firsthand. The combination of AI and materials science is already transforming how companies approach some of their most critical challenges. And much like that perfect cake, the goal is to finally have the recipe and to make it work, consistently, at scale.