The intersection of physics and AI is leading to groundbreaking innovations across various industries.
- Startups in Europe are utilising AI to enhance the design of complex products such as cars, planes, and semiconductors.
- Monumo, a Cambridge-based company, is pioneering the design of more efficient electric motors with significant environmental benefits.
- Physics-based AI models offer precision and efficiency without the pitfalls of traditional language-based models.
- Investors are urged to consider the environmental impact of AI developments, promoting sustainable innovations.
The confluence of artificial intelligence and physics is catalysing transformative advancements within several engineering sectors. By deploying AI in novel ways, European startups are poised to make substantial improvements in designing complex products like vehicles and semiconductors. Unlike their language-based counterparts, these AI models utilise mathematical and physical principles, eschewing the inaccuracies often associated with large language models (LLMs).
Monumo, founded in 2021 in Cambridge, exemplifies the potential of using AI to improve engineering processes. By developing AI-driven models to enhance electric motor efficiency, Monumo aims to make a significant dent in global carbon emissions. According to founder Dominic Vergine, a mere 5% enhancement in motor efficiency could match the carbon reductions of eliminating Germany and France’s emissions combined. This ambitious endeavour underscores the appeal of optimising electric motors, given their substantial role in global energy consumption.
Vergine explains that the strategic use of machine learning to design holistic motor systems rather than disparate components is yielding remarkable performance gains. Monumo’s engagement with over 35 companies indicates burgeoning interest and confidence in their approach. Their collaborations include discussions with a leading automotive supplier and potential partnerships with Japanese car manufacturers, illustrating the transformative potential of their innovations.
Besides automotive applications, Monumo’s methodologies are extendable to electricity generation sectors. Considering that about 95% of global electricity depends on turbines, their AI model offers a promising avenue for widespread energy optimisation. Vergine posits that replacing the traditional $5 trillion-a-year oil industry could significantly advance global decarbonisation efforts.
PhysicsX, another European entrant, received a $32 million investment to leverage AI in optimising component designs across automotive, aerospace, and semiconductor industries. CEO Jacomo Corbo highlights the superiority of physics-trained models over traditional simulation methods, delivering enhanced precision at reduced costs. This capability enables rapid testing of component designs, significantly outperforming prior methodologies and opening new avenues for precision engineering.
The discussion around AI’s future often revolves around balancing technological progression with environmental responsibility. Physics-based AI models present an energy-efficient alternative to GenAI models, which demand immense computational resources. Dominic Vergine of Monumo urges investors to focus on sustainable AI projects that do not aggravate environmental challenges. The question of prioritising AI’s ecological footprint is becoming increasingly relevant as the industry progresses.
The evolution of AI, intertwined with physics, presents a promising frontier for engineering, urging a mindful approach to sustainability.
