GAFAM Empire is a project in which we look at all the known acquisitions conducted by five big tech companies: Google, Amazon, Facebook, Apple, and Microsoft, from the moment they made their first one to the end of summer of 2022, when we stopped collecting data. Why do we call it an empire? Our focus is on trying to understand how these companies purchase their power and ability to grow their respective market positions based on their information infrastructure through the simple operation of acquiring other businesses and their products. Our understanding of empire here is very basic: empire controls three basic layers of knowledge: 1) collection of data/intelligence, 2) storage of that intelligence as well as 3) capacity of processing it. These three unique characteristics give them powerful and unprecedented insights into how our societies operate. Why are we not calling them monopolies? Because this framing based on the idea of limited resources and competition for consumers frankly does not apply - the resources are unlimited - and that is our data and attention. When it comes to competition we are free to chose the ones we like - we can use products of them all in fact. In short we call them empires in a collonial sense, as in digital colonialism and not monopolies in a market competition sense. All of these five companies dominate the space of collecting data (number of users), managing and processing that data (cloud and other infrastructures they own, including hardware) as well as processing and analysing that data (machine learning, AI, algorithms). They all monetize this dominant position, making them the richest tech companies out there. How big are they, which sectors do they dominate and in which do they compete, what futures are they trying to prepare for? It is obvious that for them the ideal future is the one where we do everything online, everything is digital and goes though their channels. These were the questions driving our research and this visualization. You can learn more about the data used here in the Methodology section, however, we have to say that the data we acquired has significant limits. One would hope that there is a clear way of learning how much each of these acquisitions was worth; sadly that is not the case. This specific data is sporadic and often unreliable. It would be great to know the exact reasons for each of the acquisitions - this is also very arbitrary and ephemeral - as the descriptions of purchased companies and products often do not reflect the actual reasons of purchases: was it buying a product, buying out a competitor, absorbing talent, skills, or expertise, or to exploit the content (which is why GitHub users filed a class-action lawsuit against Microsoft for training an AI tool with their code)? So what is it that you can actually learn here? We can learn how they grew - which sectors they expanded to, what types of know-how they absorbed, and in fact who they actually are besides who we think they are. We, the users of their services, often associate who they are with the most popular services we use, but looking at their acquisitions we are looking at a very different picture - a picture dominated by strategic purchases and political visions. In 2021 alone it is estimated that the technology acquisition market exceeded over three trillion US dollars. It is not insignificant what they acquire and we should know more about why.
In the history of big tech companies, each company that has been acquired is represented here by a single dot. Each company is described by up to 7 tags related to what they do. Self representation of companies is an important proxy to read big tech companies' interests in why they acquired these specific companies. Each company is described using different tags to define similar aspects, others in a generic way and others very specific. How can we recognize the main areas where these companies are located if we do not have a general unit to observe them? How can we observe them while maintaining the gray zones that they themselves circumscribe according to their descriptions? To analyze all the tags and give them a more general classification, the initial 390 different labels have been reduced to a total number of 25 with a maximum of three per company. From the combinations of these tags it was possible to create a proximity map where companies close to each other are described by similar tags. As a result we saw for the first time the areas that the companies defines as a landscape of acquisitions.
