Why would someone in a non-software/internet business acquire an artificial intelligence (AI) startup? AKA How an AI startup may find a niche
This article was originally published on Medium, and scored 242 views.
Even though investing in artificial intelligence (AI) startups is a recent trend, some large exits already have taken place. For example, Intel paid $15,3B for Mobileye and Google acquired DeepMind for $500M+. Recent acquisitions (e.g. Blue River Technology by John Deere and Delta ID by Fingerprint Cards) suggest that the pace of deal-making activity in AI is not cooling down.
An interesting fact is that an ongoing hunt for AI startups is also unfolding in industries one could hardly suspect in a romance with AI five years ago. Non-software/internet companies, for example involved in medical device manufacturing, petroleum contract drilling or cosmetic manufacturing acquire AI startups. Almost 10% of AI startups acquired between 2011 and 2018 had a non-software/internet buyer, compared with 3% of those scooped between 2000 and 2005. Therefore, it looks reasonable to extend my previous research that covered reasons behind largest AI acquisitions and study deals that were led by non-software/internet buyers.
Thinking about a non-software/internet company acquiring an AI startup one may apply four lenses:
- Technology/product — complementarity of an AI startup to an existing product. For example, an algorithm that complements a barcode scanner;
- Customer — an ability of an AI startup to help customers of a corporation to extract more value from its existing products. Imagine a knee implants manufacturer, who acquires a robotic startups to make installation of its implants (surgeries) easier;
- Process optimisation — when a corporation uses an AI startup to streamline its internal processes, for example to improve drilling process;
- Transformation — think of AI that disrupts industries, say by bringing self-driving cars to life.
For exemplary purposes, transactions covered in the research are clustered around these four segment (see slides below). However, they are not mutually-exclusive, and probably overlapping — it’s just that one of them is not enough to build a sustainable company that corporations would hunt for.