Our lives today are shaped by a handful of omnipresent players. Google determines our access to knowledge, Amazon governs our purchases, and Meta orchestrates our social interactions. Behind these giants, the same “winner takes all” logic dominates. Joëlle Toledano, professor emeritus and member of the National Digital Council (CNNUM), deciphered this dynamic in her book “GAFA, reprenons le pouvoir en 2020” (GAFA, let’s take back the power in 2020). Charles Thibout, who has a doctorate in political science and is an associate researcher at the Institute for International and Strategic Relations (IRIS), has focused on the specific case of Google in France, having devoted his thesis, defended in October 2024, to this subject.
How have Big Tech companies come to dominate our lives today ?
Joëlle Toledano. These players began by establishing their dominance over services that won public support, such as Google with its search engine or the social network Facebook. An economy of fixed costs, growth in these digital services that economists call “platforms” can be deployed with proportionally limited costs and powerful network effects, making the service increasingly attractive. Each new user makes the service more attractive, the targeted advertising more interesting and the competition more difficult. Sales and profits increase with data fed into algorithms. As a result, an almost inescapable monopoly gradually sets in, leaving little room for viable alternatives. Admittedly, TikTok has made its mark, and OpenAI and its competitors are trying to take Google’s place, but to date, Meta and Google still dominate to a very large extent, not only in terms of usage, but also by stealing the advertising markets that serve as ‘cash cows’ to prepare for the future.
How did we become dependent on Big Tech ?
Their products have become indispensable to the point of redefining the way we communicate, get information and even the way we consume. However, these companies were not born out of a need previously expressed by their users, but rather out of the desire to create a new need. Today, cutting yourself off from social media could have both personal and professional consequences. So, let’s not even talk about getting away from a tool as useful as the Google search engine. All the same, we must be careful not to take an entirely negative view of these companies [Editor’s note : the economic and political influence due to the market power of these companies]. Because their profits are also significant, which shows that there was a real need for them.
However, it is not enough to create a need that revolutionises the lives of users to ensure a substantial economic income. Each of these players seeks to create a world that you would have no desire or interest in leaving. It is an economy that revolves around the user’s attention and available brain time, so that they consume advertising.
In addition, the indispensability of a commercial service, such as Amazon, also leaves very little choice for traders wishing to take advantage of it. The platform, in its terms of use, has established a number of rules to be followed. The e‑commerce giant therefore reserves the right, via its algorithms, to decide how other traders access their service. The economic advantages of having a business on this site are such that, for some, complying with the rules and purchasing Amazon’s logistics services or advertising can become a necessity. The company thus becomes the total master of the commercial market, and also of part of the local economy of the countries in which it has established itself. And speaking of Amazon, their strength is also due to their AWS cloud service. There are a significant number of companies that depend on it today. As is the case with the number of economic players in each country that depend on networks for their marketing communications.
How did a company like Google manage to gradually establish itself in France ?
Charles Thibout. Relations between states and multinationals have always existed, even if they fluctuate. There has never really been a clear split between the public and private spheres. Power is never totally in the hands of a particular institution or actor. It will always be the result of negotiations between different actors. And, depending on the historical context, certain types of companies, depending on their own strengths, will have an advantage over others in these negotiations. Today, in the digital age, the web giants inevitably have more weight.
Google’s establishment in France was still fraught with difficulties. From 2003, a year after the creation of its Parisian subsidiary, Google suffered attacks from various French economic sectors. In fact, its arrival unsettled many people, and the political authorities quickly became involved. Notably in 2005, with Jacques Chirac trying to initiate European projects aimed at competing with the American giant. This project was not successful, but it demonstrated France’s desire to safeguard its national sovereignty. French mistrust of Google would eventually fade from 2010 onwards, and its image would shift towards that of a potential partner in public policy. It was at that point that Google would be seen, in politics, as a means of showcasing the ability to intervene and act on the world, even though all the indicators of political impotence were there. We see this, for example, with the use of Google technologies by the tax authorities. The perception of the multinational is changing, French politicians are now seeking to attract these giants to France and convert their capital into political resources.
Since then, has Google established itself as a French economic player ?
CT. Under Hollande’s five-year term, a slight dip could still be observed. In addition, a tax search of the multinational’s Paris offices took place in 2016. This event, justified by a tax adjustment procedure, was experienced by employees as a real attack, with particularly damaging effects on the image of their company. Google is the first company in France to undergo such a procedure on this scale. Its reaction was to do everything it could to reappear as a normal, responsible company. And, as such, to be able to benefit from the same privileges as large French companies.
A recent Stanford report highlighted that around 70% of PhDs specialising in AI end up in the private sector, and only 20% in academia.
Google then decided to become French through a process called the “indigenisation” of the company. This process involves two things. First, it is necessary to recruit French staff, but also staff in high positions in the political and administrative field, such as former senior civil servants. Then, it will be necessary to build a system of alliances with other French economic players. These alliances would become a major element of Google’s indigenisation strategy during the 2010s, once the recruitment of senior civil servants showed all its limitations in terms of strengthening ties with the political authorities. In 2013, Google obtained a seat on the board of directors of the Parisian section of MEDEF : an anchoring in the field of employer representation that reflected the growing influence of the company, and more generally of digital technologies, in the economic model of French companies.
The company has therefore established itself as a French economic player which, through the services it offers, helps the development of other French companies. In addition to becoming a “French company”, it has become a central player in the national economic field.
What is the relationship between Big Tech and institutional power ?
JT. This vitalness, granting such economic power to these giants, even extends to political communication. The Australian law debate in 2021 is a good example. Australia wanted Google and Meta to pay the country’s press better. This provoked a standoff between the two web giants on the one hand, and the country’s government on the other. Although this tug-of-war failed to get the bill amended, and it became law the following year, several elements emerged from the conflict. To signal its opposition to the initial terms of the bill, Facebook went so far as to block Australians’ access not only to news on its platform, but also to government sites providing sensitive information to Australians (relief, etc.). Essentially, the public authorities depended on this network. The result was that ministries found themselves without communication channels. In the end, Australia did pass a law, but only after making changes that made it more acceptable to Meta and Google.
These companies also have significant influence in the field of digital research and development, don’t they ?
It is true that another source of influence, and not an insignificant one, has to do with research. The total budget for R&D in France, including both public and private funds, is €60bn. For the GAFAM, it is $200bn – almost exclusively for digital technology. The vast majority of AI research spending today comes from these companies, and with this kind of budget, they can also afford to set the direction for future research. This power also implies something else in the research world. A recent Stanford report highlighted that around 70% of PhDs specialising in AI end up in the private sector, and only 20% in academia. Ten years earlier, it was 40% in the private sector and 40% in universities.
The phenomenon of dependence is therefore likely to recur for future innovations, such as generative AI. It is all based on a desire for their economic power to last, to continue to grow and to stay ahead of the competition ; the rest is just collateral damage. And it’s not a question of nationality. It’s a safe bet that if a French company had similar power, with the same regulations, its objectives would not be any different. What needs to be challenged is the business model of these online platforms. In my book, I propose a regulatory model that simultaneously tackles the market power of the dominant digital players while dealing with the problems of content, be it online commerce or social networks. This is an essential condition for regaining power. But of course it’s only a precondition for developing our own systems – it won’t happen overnight.
In conclusion, and further to the comments of our researchers, we should mention generative AI, which can be seen as the next major challenge : will we have the capacity to develop our own model, as the Chinese have done with DeepSeek, or will history simply repeat itself ?
