BurdaPrincipal Investments, has expanded its investments into the fintech space and participated in the current funding round of the saving and investing app Moneybox.
In the first half of the year, no topic was more dominant across all industries than artificial intelligence (AI). And the hype continues. What is particularly exciting in this area are the many young companies and the new business models and products they are developing.
Investment Manager Friedrich von Wulffen has been involved with AI since his student days. He is one of the AI experts at BurdaPrincipal Investments (BPI), who contributed to the latest "German AI Startup Landscape" from the appliedAI Institute for Europe – and the results of the study are impressive: the number of German AI startups has increased massively compared to last year. But how does this compare internationally? In an interview, Friedrich von Wulffen puts Germany in context of other markets, he explains how Burda's growth investor views the developments and what conclusions the venture capital (VC) firm draws from them.
Interestingly AI is not a new topic. In fact, AI has been a proper field of research since 1956. Even the currently “hyped” model architectures – the so-called “Transformers” – were already incepted in 2017, and OpenAI's GPT-3 was published three years ago. Over the last years, the so-called Foundation Models have been getting bigger and better. What's completely new since last year, however, is that the topic of generative AI and its potential has been raised to a whole new level of awareness and is now of interest to those outside the tech bubble. This is due to the market introduction and viral spread of text-to-image models, such as Midjourney, and large language models (LLM), such as GPT-4. Chat GPT, which is powered by GPT-4 and GPT-3.5, is the talk of the town. At BPI, we conducted our first deep dive into machine learning (ML) in 2018.
We are seeing a kind of "productisation" or "industrialisation" of AI through the general-purpose, large-scale foundation models: on the one hand, completely new end-user applications are emerging, and on the other hand, existing products are being enriched by AI – without the need for expensive and well-trained ML talent. We are keeping an eye on both trends, but we are staying true to our core investment hypotheses. This means that we are continuing to focus on digital products that solve real problems directly at the consumer or enterprise level and have already proven their product-market fit. New and exciting technological leaps can also sometimes lead to interesting tech products in the search for problems to solve.
We are still at the beginning of this wave. Right now, a lot of the investment is flowing into tools and core infrastructure. Most of the economic value is being captured by the suppliers of computing power. In the long term, we are mainly interested in the new products and business models that can build sustainable business models on top of this newly emerging platform.
For startups and VC investors, it is also important to note that most incumbents also have generative AI on their radar. Also outside of the Big Tech companies, enterprises are investing heavily into generative AI experiments and many already have projects in production. Across industries, large incumbents have an advantage when it comes to creating valuable AI-powered products, due to their existing customer-interfaces and -data. This can be a key competitive advantage, especially in areas where incumbents have deep, proprietary data sets and can use them to train more accurate AI models. This advantage will be difficult for small new players to overcome. Startups should focus on segments where they are not in direct competition with large (tech) players that have this data advantage. They should find innovative ways that transport the value of generative AI foundation models to end users through novel interfaces that cannot be offered by established companies without hurting their existing products.
As hinted at earlier, one major trend is that a lot of venture capital is being invested in the underlying platform of generative AI. Tools are being created and funded that make generative AI models usable for specific use cases. Good news for the investors of these startups is that established enterprise software players are acquiring ML expertise to keep up with the rapid developments. Most recently, Databricks acquired MosaicML, a startup founded in 2020, for a whopping $1.3 billion. Another major trend is the emergence and funding of private research labs that are at the frontier of foundation model R&D. This includes for example a four-week-old French startup that recently closed a €100 million seed round, largest European seed round ever!
However, we are also seeing emerging investment activity on the application layer being built on top of this platform. So far, most of the money has gone into solutions for horizontal business functions such as marketing, sales, and customer support. However, I find products that specialise in specific verticals particularly exciting. We are seeing a lot of funding in the biotech sector, such as protein discovery, and in healthcare, e.g., patient administration. We will also see more and more activity in verticals, such as education, entertainment, finance and legal, which we at BPI have been looking at for some time.
Unfortunately, Europe is clearly lagging behind the US and China when it comes to AI. Most LLMs come from the US and China. However, we also see strong players here, such as Aleph Alpha from Germany or Stability AI from the UK. What also gives me hope is that a lot of AI talent is being trained in Germany, for example at the TU Munich or the Karlsruhe Institute of Technology. There is definitely potential here.
What surprised me the most was the explosion in the number of German AI startups since 2020 and 2021. On the one hand, this shows me that the topic is currently experiencing an enormous venture capital hype. On the other hand, it may already be a sign of shortened development cycles in the AI field and the potential to build new innovations on this platform.
According to a recent study by our friends at Earlybird, the TU Munich has produced the most AI founders in Europe so far. This shows us that we have a strong AI talent pool in this country. Now we have to keep this talent in Germany! Companies like DeepL or Aleph Alpha show that well-funded and internationally known players in the generative AI segment can also emerge in this country. However, when it comes to the number of generative AI startups, the UK are clearly ahead of us. At the beginning of the year, Sifted counted 55 generative AI startups from the UK. This puts them in first place by a wide margin, with Germany in second place with (comparatively) only 17 Gen AI startups.
