At Pentech, we like to invest in companies with a product delivering clear customer benefit, pointing at a large market, and in teams who have the drive and determination to make their product and company the category leader in their chosen sector. We also like to invest in companies that have a clear competitive advantage, whether network, specialist market knowledge and experience within the team and/or, where the product has a clear technical advantage creating a significant barrier to entry. One area we are particularly interested in is in the application of machine learning (ML) to particular market segments. Our belief is that significant efficiencies will be gained via the application of this technology to established and new markets. To date, we have invested in machine learning platforms focused on drug discovery, genomic processing, construction, insurance and marketing automation. MonolithAI, our latest investment, has an ML platform focused on next-generation computer aided engineering of physical product design.
Today, all complex mass-produced physical products from aeroplanes, to the system you are reading this on, to shampoo bottles use computer aided design (CAD) and computer aided engineering (CAE) processes and tools to optimise various aspects of their design. This process encompasses computer simulation, validation and the optimisation of the design and typically includes advanced techniques such as finite element analysis (FEA), computational fluid dynamics (CFD) and multibody dynamics (MDB). With complex products, the overall design process is highly iterative and expensive, including the construction of multiple prototypes to help validate the design and engineering process. Often, issues with designs are only discovered late in the engineering process, a problem only made worse as designs become more complex with the use of such things as new light-weight composite materials. The overall CAD/CAE process generates a huge amount of data. From the computer aided design data itself, to the thousands of simulations that are carried out on each design, or slight design modification, to the data that comes from the prototypes and product in the field.
Given that new designs are very often evolutions of previous ones, it is not hard to imagine that the data that has been gathered over the years could be put to good use. Imagine if an aircraft OEM was asked by a customer to change the configuration of a freight plane by adding additional parcel shelves. Today, assessment of these changes would require a large engineering effort and months of mathematical modelling and simulation. But, aircraft OEMs have a wealth of historical data on weight distribution from the original and similar aircraft designs which can be used to train an appropriate ML model, which can help give the answer in minutes versus months. Or, imagine if a car designer could use the data they have from design, simulations, wind tunnel tests and drive testing from previous models to more accurately predict the performance of future designs of that model. In theory, most of the early proto-type construction could be removed from the design process and, instead, they could rely on an AI platform to guide them on where to focus their design and engineering efforts
The MonolithAI platform does all these things, enabling engineering companies to bring product to market faster, and for less cost, by using historical data to train machine learning models, leading to drastic reductions to the time required to be spent on simulation and design validation. We were attracted to the company by the disruptive nature of their approach, the size of the market opportunity and, more importantly, by the technical strength of the team. Richard Ahlfeld, the founder and CEO, has a long list of academic achievements including the award for the greatest contribution to Imperial College London by a post-doctoral student in 2016. He has assembled an incredible team of engineers, data scientists and software engineers, all focused on building a product that reimagines the product engineering process. We are excited to be joining Richard and the team on their journey.