Significant innovations often emerge from ecosystems. Historical examples for this phenomenon are not difficult to find. Florence, during the renaissance, produced great art and architecture, among many other inventions, and set the course for modernity. Many breakthrough inventions such as the transistor, the laser and information theory originated from the research environment of Bell Laboratories. They enabled the digital revolution as well as the rise of Silicon Valley. Whenever a great series of innovations occurs, it is usually fostered by an ecosystem. Innovation by individuals rarely happens – it usually is a team effort.
Regarding Artificial Intelligence(AI), innovators face an even stronger reliance on a well set up network. This dependency is due to the difficulty of bringing talent from different fields together, high demands on training data and slower growth cycles. Therefore, the relevancy of ecosystems was on our mind when we started building the Merantix ecosystem back in 2016.
AI in Europe, and especially Germany, has excellent potential, but it also faces particular challenges that individuals can rarely overcome on their own. Young AI companies in Germany can rely on a flourishing research landscape, many talented people and a vital industry. But they also face the challenge of bringing all the bits and pieces together in a productive manner, which requires connections to different sectors and professional networks and funding patterns tailored explicitly to AI. With Merantix, our goal is to provide an ecosystem that functions as a greenhouse for innovation and helps Founders to bring out the extraordinary potential of AI.
In Germany, we are fortunate to find one of the best research environments for AI in the world. In 2018 Germany already had 114 AI researchers per million, while the USA had 87 and China had 13. The Cyber Valley, for instance, is one of the most extensive research cooperations in Europe formed in the southwest of Germany and brings together established industries and excellent technical universities to develop new technologies empowered by AI. Moreover, with the Max Planck Society for Intelligent Systems and the Fraunhofer Institute for Cognitive Systems, two of the top academic institutions in AI are located in Germany. Overall, Germany is one of the most productive countries in AI Research.
Likewise, the German industry provides various potential partners willing to transform their respective fields using AI. Germany's automotive, pharmaceutical, and energy sectors enjoy a high reputation internationally. Those industries offer great potential for AI to help create the new era of autonomous driving, a new generation of medicine and the green transformation. Moreover, the “German Mittelstand” manufacturers who are at the top of their respective fields build the backbone of the German industry. In conclusion, Germany provides a diverse industrial landscape, ultimately offering great potential for innovators to integrate new AI companies.
Put together, the German industry and research landscape attract and produce many great talents in AI and business. The German university system was already known for generating highly skilled engineers during the past years. However, German employers often had difficulties recruiting these high potentials. With the rising relevance of AI to the German industry and research, the tendency of an AI-brain-drain slowly but surely gets turned around. Consequently, AI companies willing to transform industries are in a good spot to attract talent.
The advantageous conditions for AI described above beg the question of why AI's potential remains unused. The answer is that there are particular challenges for AI that make it quite hard to unleash a fundamental transformation of industries. Without the right ecosystem, it is hard for Founders to deal with these challenges alone.
First and foremost, we must acknowledge that Machine Learning is a general-purpose technology. This is excellent news on the one hand because it means that we can use the technology in almost every industry and for various use cases. On the other hand, it poses an additional requirement for entrepreneurs since they cannot rely on Machine Learning as a stand-alone technology to carry out their business. To build a successful venture, a deep understanding of AI is required and of a relevant application field. This demanding situation leads to what Techcrunch's Danny Crichton described as the "dual PhD problem".
Crichton suspects that startups today must have an in-depth understanding of more than one field to develop breakthrough innovation. According to Crichton, many low hanging fruits have already been picked by previous generations of entrepreneurs. Companies using cutting edge Machine Learning are not commonly built by some college drop-outs. For aspiring innovators in AI, this makes it much harder to find a team that entails the different competencies and can communicate effectively. Without a deep-rooted network that provides Founders with access to talents from various backgrounds, it will be hard for Founders to overcome this barrier.
These difficulties of bringing together the right people to build the future of AI are core challenges we aim to address with the Merantix ecosystem. When Merantix was founded, we already had deep connections to talent pools worldwide, which we contoniounly extend horizontally and vertically. Meanwhile, we strengthened our network through a very selective hiring process during which we do not only look for exceptional talent but also excellent team fit.
Founders applying to Merantix go through different interviews and two rigorous on-side-days before becoming part of our ecosystem. Once they got integrated into our network, we provide coaching and personal mentorship to help the Founders fully reach their potential. Moreover, Founders get access to our talent pool and hiring brand, giving them the best possible opportunity to address the problem of talent scarcity. Applying this diligent approach to acquire and nurture talent, we are steadily growing our team that now consists of more than 120+ members with more than 30 nationalities.
Another barrier AI startups face is the relevancy of data. A Machine Learning model requires much data to get trained. However, a mere quantity of data is not enough. The quality of the data is essential as well. For an AI to solve a particular problem, it needs sufficient data points that cover each variation of the problem. Usually, the edge cases are the most difficult instances of the problem because data for edge cases are, by definition, tough to come by.
Thereby, entrepreneurs must keep in mind that society and regulators are much less forgiving when an algorithm makes mistakes, compared to a human. These circumstances pose a problem that is not easy to solve. Companies either have to collect data on their own or otherwise are dependent on a strong collaborative network. Startups, therefore, need close ties to industry partners and regulators to develop AI and get it certified.
To build strong ties to the German AI community and give AI drivers a voice, Merantix is a co-founder and board member of the German AI Association. The Association includes more than 350 startups, experts and small and medium-sized companies. Together we share knowledge and best practices and keep close contact with regulators. These connections make it much easier for our ventures to develop innovative products applying cutting edge research.
Moreover, with Merantix Momentum, we founded an AI solutions provider that helps German corporates, SMEs and governmental institutions to employ AI in their products and processes. We, therefore, have the unique possibility to build close relationships with industry partners, opening up the opportunity for data collaborations.
Vara, for instance, our first venture, built an AI that detects breast cancer on mammography. To receive high-quality data from mammography screenings, they were able to land a collaboration with the Charité, Europe's largest university hospital.
A further problem concerns the funding of AI startups due to lower growth margins, especially in the beginning. The slower growth is due to the reduced scalability of AI products, compared to traditional software solutions. It is no surprise that we cannot fully automate a process immediately, using AI. Since we take safety seriously, we cannot just hand over an entire task to the algorithm.
Instead, we need to have a human-in-the-loop interacting with the machine for a long time as we slowly hand over more and more of the process to the algorithm. This peculiarity leads to lower growth margins than we are used to from other business models. The slower growth, in the beginning, is not a problem in and of itself. But the lack of hypergrowth early on puts AI ventures in a position requiring investment patterns different from those we know from consumer internet firms. This property of AI-products leads to a lack of AI-funding in Europe, especially since there is still a gap in European Venture Capital, in general, to start with.
At Merantix, we addressed this problem by raising the Merantix Fund of 35 million Euros. That way, we make sure to finance our ventures during their pre-seed and seed stage through to Series A. Moreover, we provide our ventures with workspaces and take care of administrative and financial tasks.
That way, our Founders can focus entirely on developing a product and bringing it to the market. Once the time is right, we help our ventures to find the right lead investors for their Series A and subsequent rounds. Our worldwide network allows us to introduce our Founders to the right investors at the right time. We believe that funding should not be a deal-breaker for innovative ideas. Hence, we make sure of this for our ventures
In the end, building the future of AI is a complex challenge, with new technologies and possibilities arising each day. Thus, if we want to transfer AI research into practice at a large scale, we need a solid ecosystem that provides access to funding, talent and data. This ecosystem, precisely, is what we are building with Merantix.
Over the last years, we made the first steps necessary to make Germany and Europe the leading environment to create innovative AI technology. In 2021 we opened the Merantix AI Campus in Berlin to be the physical manifestation of the Merantix Ecosystem. At 5400 square meters, spread across two floors, we provide working spaces for our ventures.
Moreover, we share the office spaces and our large common area with other startups, research centres, corporate innovation hubs, and investors. We gather the forefront of AI to share workspaces and ideas, knowledge, and the occasional beer. That way, we create a unique, intellectually loaded business and research environment at the heart of Europe. During the next years we plan on expanding our ecosystem and continue leading the way for transferring research into the industry.
Tap into our network, and start building the next leading AI venture with us as a Founder or an early stage team member. Apply now: merantix.com/careeers