The direct complicity of Big Tech firms in the genocide in Gaza, combined with their explicit rhetoric shift to the right atop a second Trumpian wave, has dented their carefully crafted public image of being progressive, green, benevolent facilitators of technological progress. Their more explicit interference in global politics, too — most recently exemplified by Meta’s role in hindering pro-Lula accounts during the elections in Brazil — has pushed countries outside the United States to urgently take stock of their dependency on the infrastructures provided by these firms. Such behaviour on the part of Big Tech is far from novel; yet its sheer brazenness has sparked a resurgence in efforts to analyse the source and nature of the power they possess in our societies, and to explore strategies to undermine it.

Argentine economist Cecilia Rikap has engaged with precisely these questions over the course of the past decade. Her work attempts to address questions around how knowledge is produced and co-opted under contemporary capitalism, resulting in empirically rich studies of how firms today act as intellectual monopolies. Through studies located across the tech sector, healthcare, energy, and logistics, her work traces how intellectual monopolies — of which Big Tech firms are paradigmatic — assert unrivalled power over the economic, the social, and the political alike. She has authored a number of important books on the topic; her work has also appeared in a wide range of academic journals, as well as in public venues such as New Left Review, Phenomenal World and The BREAK—DOWN.

In a conversation with Disjunctions editor Vinit Ravishankar, Cecilia Rikap talks about her most recent book — The Rulers: Corporate Power in the Age of AI and the Cloud, published by Verso in August 2026.

---

In The Rulers, you draw a distinction between firms that control chokepoints, and those that resemble panopticons. This is a really interesting duality, but there are a lot of firms that don’t fit into either category, yet draw large incomes through their ownership of intellectual property — pharmaceutical firms, or Nvidia, for instance.

Nvidia doesn’t have both a chokepoint and a panopticon, it just has the former. For me, what’s relevant is allowing ourselves the time to reflect about the specifics of corporate power. Otherwise, it becomes very hard to identify the toolbox that we’re going to use to fight against it and to build alternatives. In economics, there is this idea of the monopoly: the only company in the market, exercising power over consumers in that market. I’m not saying that we can’t see this dynamic in specific markets today, but a much more complex one also exists — especially when we stop viewing capitalism as an aggregation of independent markets, but rather a way of producing and exchanging and consuming in an interrelated fashion.

So even if we want to think at the level of markets, all markets are interconnected, forcing us to think of corporate power differently. And if we think in terms of networks — how do you obtain power in networks? Either you become a bottleneck or a chokepoint, because many other actors may need to pass through you to do something, or get something, or produce something. Or you have oversight, and act as a panopticon over the network, which even allows you to possibly steer the trajectory of the network. These are not necessarily the same type of behaviour.

In a way, it is more illustrative to start distinguishing between companies in terms of the dynamics of how they exercise power. There is also the question of the source of that power, of course, but because of the focus of The Rulers, I was less interested in firms that benefit from forms of power that emerge from the extractivism of natural resources, like oil companies. Tech firms do indirectly engage in this themselves of course, through building data centres and so on — but this isn’t the basis of their power, more of a collateral consequence. The explanation for their capacity to control beyond the ambit of their ownership comes from the fact that they can systematically capture and turn different forms of knowledge into assets that they possess, that in turn enable sustained value capture.

So for me, this double distinction is useful. Going back to your point about Nvidia — indeed, Nvidia does have a huge chokepoint when it comes to tech, and digital technologies like AI do seem to be everywhere these days. Regardless of what we do in our personal lives, political lives, in our workplaces, organisations, the public sector as well — tech underlies everything. So this chokepoint can indeed be massive, but it is still just a chokepoint. Since both Nvidia’s business and its intellectual monopoly are very concentrated, it is also to some extent dependent on the success of AI adoption today. Companies with both a panopticon and a chokepoint, on the other hand, like the cloud giants, they not only become an obligatory passage, but also have this oversight over the system, this access to information that also enables them to steer the trajectory of the system in directions that reinforce their power.

When reading about this — and this is explicitly referenced in your book — I am reminded of Baran and Sweezy’s analysis of monopoly capital, from 1966.¹ It seems that the main change today has been a shift from tangible assets to intangible assets: would you also frame it that way?

I actually wouldn’t — the focus of scholars may have changed, but not the source of power. Because behind the idea of Monopoly Capital, we have something very similar to Chandler’s idea of the vertically integrated enterprise, where the advantage was that everything was centralised — therefore, the information to organise, control, and plan the value-chain was centralised. So the vertically-integrated enterprise was already an intellectual monopoly, but one that monetised in-house produced knowledge. We see this quite clearly in the work of other Marxists like Harry Braverman, whose work explicitly explains how the monopolisation over technology and over knowledge enables bosses to squeeze more from workers.

And so for me, the change from the age of monopoly capital to today’s moment is the capacity of the leading corporation, of monopoly capital — whatever you want to call it — is the ability to control beyond ownership. In the past, firms needed to be vertically integrated to control workers and capture their workplace knowledge; today, they can outsource a lot, including the production of new knowledge, while continuing to assert control over workers.

Braverman is a great example to bring up in light of all the discourse around profitability in the context of AI.² Do you see this as another Braverman moment — of de-skilling workers and transferring control to management?

This is quite important when we discuss what type of AI we are getting. At its present trajectory, this is clearly a technology that contributes to the de-skilling of that portion of the working class whose jobs involve thinking and doing. This is also why people working at a factory will not see the effect of generative AI in their workplace. They might see it in their personal lives, but not in their workplace tasks, because those have already been altered, already been detached: workers in factories already do not have control over their work process. People working as journalists, as academics, as lawyers, or as designers or software developers, on the other hand, still do have at least a certain degree of control over their production process. It’s precisely these tasks that are now targeted by generative AI. And though it might seem that we all become managers of some kind now, that we all control the agents, it is actually the agents that are dictating the pace of work. And this ends up helping actual managers, not just to control us and the bottom line, but also to project control into the future, and engage in corporate planning more effectively, as labour movements are weakened through workers being de-skilled by generative AI.

I’m thinking of how in Technofeudalism, Cédric Durand describes this with the smile curve, highlighting how labour discipline is moving outwards from manufacturing.³

But I think that example was already incorrect. If you think about R&D, it’s a fantasy to think that universities are the heart of academic freedom, or that corporate R&D laboratories are places where we flourish with our ideas. Our imaginations are heavily conditioned and constrained by so many factors — the contexts in which we were raised, where we work, what we do, what sort of literature we came across, what we decided to read, and so on. And precisely one of the things that my research does is to show this: to show that this capacity to control beyond ownership, whose history we can already trace from global value chains onwards, has today been reinforced, and has increasingly expanded into the co-production of knowledge itself. It’s not just the science and the technology — even the ideas and narratives that are co-created by the many are increasingly appropriated by leading corporations, for the benefit of the few. These few are the same that dictate and influence research agendas and public discourse.

It is true that for many forms of labour, there used to be no such thing as the stopwatch, though we may well end up getting one with generative AI. Or rather, we might witness what I describe in The Rulers: some sort of transition from the stopwatch to the smartwatch, where each of us will have our own clocks according to our individual capacities, to make sure that capital can squeeze the absolute maximum from each and every individual. This is even better for capital than normalising labour to the stopwatch, where you might end up missing out on how some people can actually work more, and that you can extract more from them. Nevertheless, the discipline was already there — it was just more invisible, and not as bespoke.

Another issue with the smile curve is that it also doesn’t emphasise how the issue was never that of value creation, but of value capture. After all, who decides that R&D creates more value than manufacturing? All of this is part of how we choose to value things, and capitalism is very much connected to those making the choices.

Along parallel lines, I want to talk about the entire debate around technofeudalism. Even beyond Varoufakis, I’m reminded of the debate with Evgeny Morozov in New Left Review, almost five years ago now.⁴ How do you think that debate has held up today?

So even when I intervened in the NLR debate, what I was trying to bring to the table was that we need to understand the nuances and sophisticated ways in which corporate power is built and exercised today, and how it transforms capitalism into a society that is planned in an authoritarian fashion. This helps us both when we are thinking about political alternatives, as well as when we are trying to frame our ways of living today, and how to deal with them. What I want to say here is that ultimately, I don’t care about how people name the world we live in, as long as we make an effort to understand that there is no easy band-aid here but that it really requires us to radically rethink our societies —where we produce collectively, but remain unable to decide on what we produce, or how we produce it. And on top of that, we also need to rethink our societies with a constraint: the elephant in the room, which is planetary boundaries. And we might think that it’s unfair, that it’s our generation that needs to deal with ecological crisis — I understand. But unfortunately, two plus two makes four, and we cannot change that, at least not today.

So this would ultimately be my point about how we name this. In academia, we want to put names to things; I do use a lot of concepts myself, but ultimately, what I’m trying to show is that this society is unequal: not only at the economic level, which is obvious, and not only at the social level — which is also quite obvious, given all the forms of racial inequality, gender inequality, geographical inequalities, ecological inequalities. What I’m trying to say is that we also live in a very politically unequal society, even though we have the formal right to vote. This is not just because governments make decisions that go well beyond the scope of what they were elected to do, but also because even they are today ruled by forms of corporate power, in completely authoritarian ways that nobody decided on. Nobody voted for it. Nobody said that they wanted AI for genocide, or that they wanted AI that replaced people’s capacity to think.

I’m really interested in what countries outside the U.S. and China can do to resist this. As you’ve pointed out in The Rulers, ideals of digital sovereignty often tend to feed into digital nationalism with no real sovereignty, given that most of the world relies on U.S. clouds. What could an alternative to this look like?

For me, digital sovereignty is about democracy, about building spaces for everyone to decide — on an equal basis — what technology we should be producing and for what purposes, what data should be retrieved, what data shouldn’t be retrieved, how many data centres are okay, where they should be built, and who will operate them. This is not a question about whether they are produced inside Europe or not, or whether they are run by European companies, because it’s actually a discussion about public decision-making. It is thus also very much connected to identifying the key components of this entire ecosystem or network that need to be publicly-owned — which, incidentally, doesn’t necessarily mean “owned by the state”.

Ultimately, digital sovereignty also boils down to having the political imagination to create the institutions that can govern and run these alternative pieces of technology in service of public needs and desires, and in service of truly creating a more balanced playing field. This isn’t even necessarily something beyond capitalism: there was a time when the critical social infrastructures were genuinely publicly owned, under what would now be called state-owned enterprises. Public-owned enterprises could be a better name, since it pushes us to think beyond national borders, beyond governing parties, beyond the state as an institution.

On that basis, we can also start connecting existing initiatives — decentralised apps, like Signal, for instance. Yes, it is good that Signal exists, but what about other forms of digital technology? We can’t stop at single, specific alternatives here and there; we should instead aim to offer platforms where all these initiatives can flourish, and connect to one another. We may even have some idea of competitive markets here, since our broader society remains capitalist; but this is still better than the present, where powerful firms decide both what is offered and how it is offered.

That reminds me of yet another debate that you also briefly reference, about whether capitalism today is characterised less by planning through markets, and more as corporate planning.

Absolutely. I describe it as authoritarian planning because it’s not just corporate planning: it’s also the Chinese government planning, the U.S. government planning, and all these planners fighting one another, but also establishing coalitions when doing so is in their interests, where they make concessions to each other. All of this is also why I start The Rulers by arguing, first, that the coalition between the U.S. government and Big Tech is not novel. It’s not something that started on Inauguration Day. This ruling coalition may have different balances: today, it might seem as if firms have more control, and that the U.S. government is dependent on Big Tech firms at an unprecedented level, but this is not universally true. ICE, for instance, depends on Palantir, but so does Palantir depend on the U.S. government, who are responsible for half of their revenues. This is a kind of mutual interdependence, where it’s less clear who’s calling the shots.

But when you’re Microsoft, on the other hand — you might sign a lot of contracts with the Department of Defense/War and get a lot of money from them, but they aren’t your main source of income. What you’re getting instead, through those contracts, is a political blank cheque. You are becoming more powerful internally, not only by getting a seat at the decision-making table, but by actually designing the table, deciding what’s on the menu, and who else gets to sit there. And relevant to what I was saying earlier, this illustrates some sort of relationship between authoritarian forums of corporate and political planning, that in the U.S. end up favouring Big Tech firms more. But then, up comes the Trump administration, trying to reclaim part of this power for the state, partly because Trump sees himself as the state. This has led to an interesting scenario, where we see Trump behaving more like we would expect Xi Jinping to, while the CCP acts more and more in ways that we associate with the U.S. — such as in trying to uphold the international world order, where Xi is now in the forefront asking for deeper international cooperation.

But going back to your question: for me, step one is to show that we don’t live in a market economy. We live in a planned economy, where planning is exercised in an authoritarian way. For corporations at the very core of capitalism, this is still preferable to competition — to market economies, prices being decided in the market, planning limited to inside the firm, and so on. But the feasibility of today’s extended forms of planning really illustrates that the calculation debate is over. We need to turn the page to see that there is another debate going on, between today’s forms of authoritarian planning, and whatever it is that we can offer to counterbalance that.

That precise imaginary needs to be democratic planning, which recognises the feasibility of planning itself, but democratises it to address today’s fundamental question: the question of political inequality, of who gets to decide, how we decide, how we can make every space in our lives more democratic — the neighborhood, the workplace, and so on and so forth. But because the world is so interconnected, we cannot simply assume that bottom-up initiatives will be enough. We need to make the problem explicit, and create the institutions that can centralise planning without becoming authoritarian.

On which note, there has been a lot of recent discourse around ideas and imaginings of a “socialist AI”, or of workers taking back control of AI.⁵ Is this a useful political program?

I understand this impulse, especially in the face of all these forms of epistemic totalitarianism through which we are losing more and more ground in terms of even being able to think — since the more AI is adopted, the less people think by themselves. I understand the whole push towards saying things like: let’s hack the system; let’s try to use AI for good. This will still not be enough, however, and it still leaves us within the tech debate. And this is not a debate about tech: it is a debate about democracy, and how we have a very small window of opportunity to radically rethink what democracy means, to rethink how we build more democratic societies, how we govern our lives. This doesn’t stop at just discussing how we produce things, which makes the “let’s build AI for the people” position a bit flawed.

In Brazil, for example, the Movimento dos Trabalhadores Rurais Sem Terra — the Landless Workers’ Movement — is trying to build its own AI. What they are missing is the question mark. Do you need AI in the first place? Before jumping into producing another LLM, what are your needs? What are your problems? What are your desires? What gives you joy? What is your vision for a different life? How can we materialise that, make it happen? What changes would we need to make? What are the transformations that we envision and where to start from? Based on these questions, we could say that we might need technology for certain things. It might even be AI — but would it be generative AI? Probably not.

Because even apart from questions of planetary boundaries, general-purpose models that try to do everything end up doing things in a very non-optimal way. Even within the realm of AI, the idea of having smaller models, targeted to specific tasks, to solving problems in a more effective way and usually things that humans cannot do, is ultimately maybe something we can do more of, once we identify whether we truly need them. We can, of course, also keep some space for experimentation; and we can try to hack the system in between. But this will go nowhere until we craft an alternative strategy that, yes, understands the role of hacking the washing machine — but also understands where technology can fit into a broader strategy, and what that broader strategy is about, and how we get there.

Ultimately, it’s easy to say we want to live a life that is meaningful, and affords respect and equal opportunities to everyone, within planetary boundaries. This is very abstract. How can we design the institutions that guarantee this? How can we think about the transformations that we need to build today? How can we rebuild existing public institutions like unions or workers’ parties? How can we coordinate the interplay between these new institutions, as well as community movements — such as resistance against data centres — to arrive at some form of democratic planning of our lives?

Notes

1. 1. Paul A. Baran and Paul Sweezy, Monopoly Capital: An Essay on the American Economic and Social Order, 1966.

2. 2. Harry Braverman, Labor and Monopoly Capital: The Degradation of Work in the Twentieth Century, 1974.

3. 3. Cédric Durand, How Silicon Valley Unleashed Techno-feudalism: The Making of the Digital Economy, 2024.

4. 4. To follow the debate, see: Evgeny Morozov, “Critique of Techno-feudal Reason”, New Left Review, Jan–Apr 2022; Cédric Durand, “Scouting Capital’s Frontiers”, New Left Review, July–Aug 2022; Cecilia Rikap, “Capitalism As Usual?”, New Left Review, Jan–Feb 2023; Jodi Dean, “Same As It Ever Was?”, Sidecar, 6 May 2022.

5. 5. See, for instance: Holly Buck, “Democratic Governance of AI Is the Real Solution”, Jacobin, 26 April 2026; Evgeny Morozov, “Socialism After AI”, The Ideas Letter, 11 December 2025.