Book Launch/Lecture by Marek Poliks & Roberto Alonso Trillo
Transcription
Davide Tolfo:
I will do a very quick introduction in Italian. Then, of course, the presentation will be in English. Thank you all so much for being here. Today we’re joined by Marek Poliks and Roberto Alonso Trillo, authors of “Exocapitalism”. “Exocapitalism” is their latest book, published by Becoming, a Berlin-based publishing house. It’s a wonderful publishing house for us because, in fact, it resonates deeply with many of the texts we have here, and I must say it has also found a good audience among the people from IUAV and Venice in general who come here looking for books like this one— books that, in a way, elevate what is called “theory” in Italy to a higher level.
In fact, everyone who has stopped by has heard me, even if unintentionally, talk about this book.
Marek Poliks:
Every time we do this, I love to just acknowledge the incredibleness of Becoming Press and encourage everyone to just buy the entire catalog. We’re going to give this in the form of a lecture, because we're serious theorists, if that's okay. This is the kind of second stop of the first kind of leg of a tour. And then we have quite a big tour coming up. So we're kind of using this opportunity to kind of optimize a little bit, and we'd be very interested in feedback and perspectives. I think it's a really, really wild talk.
It's full of examples. I'm not going to read from the book or anything like that, this is all new material precisely for the tour. We've been doing a lot of media conversations about the book. And I think every time we're asked to talk about the book or given a context to talk about the book—and again, very, very grateful to Bruno, this is like one of the best bookshops I've ever been to in my life, big shout out to Bruno— every time we're given a context to talk about the book, we have a tendency to want to talk about the entire thing.
But it's like a wormhole, I mean, it really sucks hours of conversation out. So we're going to try and stick principally to exactly one concept. The book is made of a few key concepts: Scale, Fold, Lift, Drag, Lag and The Last Mile, which are all kind of alien concepts to you probably, if you haven't read it. But, they are essentially ways, or a kind of machinery that allows us to reorganize the vocabulary of finance, specifically the relationship between value and price, which might be pretty abstract for this group, but you can think of it as a kind of toolkit that helps you kind of navigate the complexity of 21st century capitalism in a way.
But, I'm only going to talk about one of these concepts, which is called Lift, which is my favorite concept from the book. And I think it's kind of a key to opening up the rest of it. So let's begin. It has been understood for a very long time that capitalism has something to do with abstraction. Marx, for example, identified a kind of formal operation in which money becomes more money through the digestion of the Earth and the subsequent metabolization of that stuff into abstract capital. And that metabolization is a kind of Lift, as we call it.
It's a phase transition from discrete physical utility, usefulness, into virtual space. The holy stomach converts our dust ultimately into energy. A commodity is an abstraction of a physical object, right? It's resituating something real into a virtual plane of exchange that's nonetheless real, but virtual. It lives as a coordinate on a graph of money. It's been lifted from a simple material form, and it's been afforded a kind of representational surplus. But we go much further than this, suggesting that it’s not only capitalism that has something to do with abstraction, we actually go really crazy here, we suggest that actually abstraction itself has something to do with capitalism, that the kernel of capitalism lives within the formal logic of representation itself, and that capitalism might be something as transcendental as language.
And that's a crazy take. For anyone new here, that's an absolutely insane thing to say about capitalism, but that's totally a hill we die on. This is too much to talk about at the moment. We're going to keep it actually very grounded in real world examples. And I'm giving you a little sneak peek as a kind of taster of where we conclude. So I’m going to stick to Lift and offer up all these examples. So if I'm to put together an elevator pitch for what Lift means, and Lift being this concept that I want to talk about today, Lift is the drive that compels capitalism towards abstraction and away from production.
And I would double down on this statement by suggesting that not only is capitalism allergic to production and consumption, it's further allergic to the feedback loop of consumption and production in general. Moreover, it's allergic to making things, selling things, buying things, distributing things and organizing things. Further, that it's allergic to any kind of physical media whatsoever, and even more importantly, that it's actively disinterested in sustaining any direct entanglements with the physical media in any respect, humans included. Period.
And I don't mean disinterested or allergic in a kind of psychoanalytical sense, because in Exocapitalism there is no psychoanalytical subject, Capitalism is not an ambient ghost powered by human psychosocial drives. Thinking about someone like Žižek here, or Italy's patron saint at the moment, Mark Fisher. Instead, when I say disinterested or allergic, I mean this strictly in an algorithmic sense. The cascading self abstraction away from material conditions isn't just what capitalism does, it's what capitalism is.
Roberto Alonso Trillo:
So, we want to start with something that is quite simple. We want to keep it simple. You can only extract so much profit from a physical commodity. You need to make it out of stuff. It needs to be formed by people. It needs to be moved around in space. All of those operations actually cost money, and the costs of those operations do not scale. They can only get so cheap. You can only, if you think about it, you know, the per unit costs have hard floors and the capacity constraints and, you know, we have economies of scale that taper to real limits.
You can buy as many raw materials as you like, but you're limited by the physical size of the component marketplace, the raw materials in the earth, mining machines, geopolitics of it all, etc. You can enslave labor as much as you like, but you're limited by local loss and ultimately by the reproducibility of the labor force. People having kids. You can optimize logistics as much as possible, but at the end of the day, there are ports, tariffs, rides and fuel that need to be paid for. And unfortunately, the price on the commodity you're trying to sell is limited by the demand. You can charge so much. And the differential between the cost of the goods you've made and the price you're selling for, that is the profit margin is limited. You're kind of stuck, if you want to put it that way, with a price, you can't make up a price. It's kind of grounded into something somehow.
If people stop buying your commodity, you're kind of fucked because you've had to source the underlying resources required to produce this commodity at the scale for which you forecasted. And so if your brand's attention tanks, or if a newer version of the thing that you're trying to sell outcompetes your product, well, you're sitting on a pile of components and assembled goods with nowhere to go, and you're paying for storage at the end of the day.
For so much theory downstream of Marx, this is the wall space of capitalism: the stubborn, naive, let's put it that way, materialists of the world are only able to imagine commodities marketplaces operating under extreme price constraints. So the only way that profit can be generated is by trying to squeeze your production costs. Essentially, at the end of the day, and this is common parlance, I think we all heard this, stealing from your workers, right? If this is the case, theory downstream of Marx interprets capitalism as a labor dynamic.
It understands capitalism exclusively through the production and consumption of commodities. And commodities face positive marginal costs and price ceilings set by demand. And of these factors, yes, we agree labor is the most apparently compressible to some minimum viable cost. And again, theory downstream of Marx interprets capitalism as a social dynamic through which capitalist and worker classes sustain a holy struggle and against which the worker class is not only subject to exploitation, but psy-op.
Deep social and psychological engineering that make the terms of exploitation seem reasonable to the worker class. And look, we are oversimplifying things quite a bit. At the very least, we can say this clearly: the left, especially again downstream of Marx, seems to get lost in the idea that this process is the only way that capital creates more of itself, that this is the way that capitalism actually grows. And we can say that the point of the book is to say that this is wrong. This is not the way that capitalism works.
This is not how capital arrives at more capital. It is totally our contention, and Marek was pointing out that this might sound crazy to you, that capital can generate more of itself in totally frictionless ways without any necessary relationship to production, consumption, or labor. And it is only in capitalism's appropriation into human scales and through human technologies that it has encountered these limits. Capitalism has been trapped in commodity form, but critically, it's been able to find its way out.
And the first distinction that we want to make is between what we call hardware and software, which we mean to be a bunch of different things, like the distinction between physical media and virtual media, the distinction between industrial production and energetic production, the distinction between B2C and B2B, and the cascading distinctions between commodities, services, managed services, and at the end of the day, partnerships and ecosystems. So again, the margins of hardware suck. They are a combination of components, labor and logistics, all of which have strong limits on how much they can scale.
But on the contrary, the margins on software are supple and elastic. At first, yes, you need software engineers to create IP, manage dependencies. You have to do research and development, but in the software business, you need this kind of labor only at the beginning. You need this as much once you've achieved a level of scale. We think that a lot of people miss this point. There's actually very little IP in software development, very little authentic new creation of stuff. This is something that everyone gets wrong, including people huge to us like Tiziana Terranova and Mckenzie Wark, who we interviewed in a couple of weeks ago, who we think misread the software economy as a permutation of, or even a one on one mapping of the commodities or hardware economy.
But, believe it or not, most software businesses are basically doing resale. They're essentially leasing IP that is out there already, packaging it together with a bit of an old IP and then selling it back into the market at a significant markup. Software engineering labor is not creative labor. It's not even intellectual labor. It's a low overhead maintenance operation that concerns itself principally with the management of integrations between different pre-existing third party software ecosystems by a new third party component as the business scales.
In turn, after you've done the initial work of scaling up a software business, the intellectual labor of software development becomes less and less important, especially against the energetic labor of raw compute power itself. And we can cite here and quote a million think-pieces on the automation of software engineering practice. But again, Exocapitalism is not excited by the prospect of facilitating labor. Instead, it is activated by the possibilities of building more and more surface area into software engineering itself through modular, low code or no code environments, and especially through GenAI tools like Coursera or GitHub, Copilot, which do not automate software engineering labor, but rather nest opportunities for Lift, which is this idea that we're trying to discuss today, inside of software engineering labor.
The limits then on margins have little to do with the intellectual property made by software engineers, let alone the completely fungible sales of back of house labor used to scale a business. Because, let's be real, every sales or back of house tech job is functionally engineered to be exchangeable across the industry. At scale, the floors of software or software margins have far more to do with compute. The bill at the end of the day is your AWS bill: the amount you pay, either directly or via the modules you've purchased and added into your environment to manage data center services, who are themselves several orders removed from the physical manufacture of computers and electricity.
Software, and this is important to us, can basically have any price. It doesn't matter. Software is dominated by fixed costs and very low marginal costs. The main variable component at scale is compute, Sure, there is a market determined ceiling on how much you can charge for software, but again software profits commonly exceed over 80% of the costs. It's imaginary. It's actually fiction. And this is what makes Exocapitalist activity, not the profit margin itself, not even the profits, but the fluidity and flexibility of price.
If there is surplus value generated in the software economy, it is not extracted from human labor, but from electricity itself. Energy is actually still today very, very cheap. It is maybe too cheap. And here we feel attracted to George Bataille, Nick Land and Reza Negarestani on the concept of the solar economy. We are not referencing here solar panels, but rather the postulation that obsolete, functionally, infinite energy as represented by nuclear fusion is understood by Exocapitalism as a given.
Another sidebar, we are extremely interested in understanding labor strictly through energetics. The idea that nuclear fusion is the apogee of labor, or that maybe labor has always been a proxy for nuclear fusion, but this is perhaps a long conversation that we can have later. Regardless, it follows, then, that the most basic kind of Lift is the movement Regardless, it follows, then, that the most basic kind of Lift is the movement away from production, consumption, and logistics of physical commodities toward assemblers, licensing and renewable capitalization of software.
Marek:
Okay, so to bring it really, really down to Earth here. We are going to look at Siemens, since we all know what Siemens is. It's like the hardware company, in the sense that the Siemens logo is everywhere. It's like the hardware company, in the sense that the Siemens logo is everywhere. It's on like toilets, wind turbines, industrial manufacturing equipment. But the point that we're making is that it has always been a software company and it is a software company, which is, again, like a really insane thing to say, but we're sticking to it.
Their stated ambition is to transform the industrial world into what they call or what everybody calls digital twins. So a digital twin is essentially a kind of 1 to 1 representation of a physical thing out there in the world into the cloud. So, there's a physical object out there on a manufacturing line or a toilet, I guess, that has a button on it. Theoretically, you could hit that button in your browser, right? You could hit it with your phone, whatever, call an API and make that thing happen.
Or if you have a piece of industrial equipment with a thermostat on it, you could look at that thermostat right by looking at your phone. So that information will be published. So that's the idea of a digital twin, it’s this 1 to 1 representation of a physical thing in a virtual space. That's something that Siemens has really been interested in for quite some time. And you can see how that might be useful in terms of automation. You can create, quote unquote, smart factory where everything gets automated and you have all these kinds of pipelines of data flowing into each other, where, you know, everything can be accessed, maintained and optimized through just digital interventions.
So you don't even need any people working in this factory. But this is a kind of meagre kind of Lift. It's a very first order virtualization of physical assets in their relationship to digital space. And I think, people really get fooled here because they think that intentions like this— like smart factories, factory automation, for example, Tesla always gets cited like this— they think that these companies are actually interested in doing manufacturing, in automating manufacturing itself. And therefore driving down the variable cost of production.
But that's actually not what's happening here. Like Siemens is not interested in manufacturing for its own sake. Instead, they're actually pivoting away from manufacturing. They're lifting from providing manufacturing goods to working with manufacturing businesses and instead of selling hardware to businesses that operate that hardware in a consumer facing way, their goal is to sell software to businesses that then, in turn, sell their manufacturing capacity to businesses that sell hardware. So in the case of Siemens, think about a wind turbine.
You can sell that physical wind turbine to an energy utility. Then they will take that asset as some kind of CapEx and they'll resell that product of the wind turbine energy to their customers. And that energy utility, you call them B to C in American “biz” parlance, which is now global “biz” parlance. They're engaging in a direct relationship to people, to C. They're the B, business, to C, to their consumers who are consuming that energy. Their customers are the end of the journey of that business. So if Siemens is selling that utility physical wind turbines, you could call them a B to B to C configuration, because they're selling physical assets to a B to C company.
But instead what they're actually doing, instead of trying to sell to that energy utility, B2C, when building digital twins, you're trying to sell to B2B, you're selling to companies who are then themselves selling those wind turbines to businesses, which only then ultimately find their way to that B2C business at the end of the line and ultimately that end customer. So Lift is the movement up the chain. So you start with B2C, you start selling some physical things to people, then you move to B, to B to C, you start selling something that a business can then sell to a person to, then B to B to B to C, you start selling something to a business that then resells it to another business, who then does something with it to produce a thing that then gets sold to a customer, right?
I know it's crazy complicated, but that's like the lowest order of what's happening in the world right now. So you're moving further and further away from the dependencies that come from the last mile production and the foreclosure of the fictional price upon its determination by the customer, by the consumer, by the last loop where there's no more resale happening, it's just being absorbed and taken by a person as done. And so what is actually happening is that Siemens is not just lifting from production, they're actually lifting from consumption from the termination point at the end of some commodity exchange lifecycle, because it is at that point when it hits the customer in a sense that the contact with shrinking ceilings and rising floors of how much you can sell something for, price determination ultimately arrives.
But Siemens doesn't stop lifting here. The goal is not to stop with digital twins, but rather to sell management software on top of digital twins. So if you're building a digital twin based on some specific piece of hardware, like let's say you build a digital twin for a toilet, you're limiting the addressable market of what you can sell that to, to the people who have that toilet. That's only a couple of people, not everyone has the specific toilet. So you're stuck with that kind of configuration. So if you're, let's say, selling a digital twin for a PCB milling machine in some kind of industrial manufacturing line, well, you're just stuck with people who can use that software, which are people who have that physical piece of software.
So that's a really annoying limitation for everybody. But so what Siemens actually did is that they started to sell hardware agnostic management software. So software that can kind of be run on anything. Siemens’ Xcelerator, which they call is available anywhere for any device, gives you the ability to sell this high margin, kind of weird, totally agnostic piece of software without any determining dependencies on anybody and what they actually own. And what's even crazier is that this product, Siemens Xcelerator, isn't actually managed by Siemens.
It's literally a third party solutions marketplace. So, this is just a 5% of the brands that are associated with Siemens Xcelerator. It's like this huge morass of like third party companies, subcontracted entities, partners, alliance members, things like that, that they're actually reselling in the form of this Xcelerator product, which then goes on top of digital twins, which then goes into manufacturing businesses, which then manufacture things that ultimately go on to end customers. You can feel this kind of movement up the chain.
And then in 2024, Siemens’ Xcelerator became ‘accelerator as a service’. If you hear as a service, you know that something is getting lifted. That's like the euphemism for lift in a lot of respects. So, software as a service, which was the original ‘as a service’, means that a transformation of software from a thing that you buy to a thing that you rent. I think we're all familiar with this and that the software itself is ultimately running on a server in your provider's back end chain, so your computer is not running the software anymore.
Your vendors are running the software, right? And in doing so, they can charge you a recurring fee for access to that server that's running a piece of software. They're not selling you access to a thing, they're actually leasing you access to a thing, right? You know, rent seeking behavior. If you're an Adobe user, like you use something like Photoshop, for example, or InDesign. You used to own InDesign, right? I remember the days we could own InDesign, like run it on your own computer. That was sick.
But now, of course you lease Photoshop, right? You have to pay them a recurring charge for the ability to, you know, own a license to this piece of software, which is kind of running on your computer but mostly running on their backend. And it's hard to think of almost any piece of software that you can buy these days that doesn't involve this kind of recurring subscription license, a browser or API driven interface. So you're not running software on your own computer anymore, right? You go to a browser, you open a page, and you do stuff in a browser now and then a license period of access.
So you sign a contract, for example, like any app you buy on the App Store, right? You sign a contract that's then renewable with some criteria, blah, blah, blah. In order to actually run something on someone else's servers somewhere else. I can think of a few exceptions. Ableton Live I imagine, Final Cut Pro, perhaps, where you can actually, you can buy the kind of unlimited license for the most part. But what's interesting is that the transition to AI driven software means that even these tools are going to grow dependencies because like once Ableton Live or Final Cut Pro turn to Runway, for example, once it becomes a company that's using AI to actually do things, it's not going to be able to be run on your computer anymore.
So even if you own a license for Final Cut Pro for some video editing software that you maybe used to own, you're going to be forced ultimately into this kind of subscription level of exchange because they have to run those AI models somewhere and they're really beefy and powerful, and they're going to live on someone else's server. So ultimately they're going to pass that price back to you and you're going to start to engage in this kind of relationship. You will never own software again, period.
What's interesting to this is that this presents an opportunity for a software business to kind of change the way that it relates to you. So instead of a single point charge from one business to a person, you can enter into a sustained long term relationship between businesses and customers and then ultimately businesses and businesses, which allows you as a business to talk about your finances, not in a way that's like, how much money do I have? How much revenue do I need to get this year from selling software to people?
But actually, how much money will I have three years from now? And I can then amortize that into how much money I theoretically have at this moment. Right now, in the software business, you never talk about your cash, like how much cash you have. You talk about your annual recurring revenue, how much software spend you have committed for the next couple of years. That's how you raise more money. That's how you reinject capital back into the business. It's like a totally speculative way of relating to the amount of money that you actually have in the bank, which weirdly doesn't matter for most software companies, only when they run out of money.
And so, you know, in the kind of first round of presentations of this talk when we got our first kind of critique of the book that basically said, you know, you're critiquing the software industry, not industry in general, but this is Siemens, Siemens has been around since Marx. Marx was alive, when Siemens was founded, right? This is like the OG hardware company and so this process of softwarization is baked into their earliest foundational movements as a business. We're going h.a.m here. We’re going crazy.
It's our allegation that all industry is and has always been software, which is a crazy thing to say, of course, but that's really what we believe in the way that it considers its relationship to revenue, and yeah, that's how it's been forever.
Roberto:
So now we’ve introduced this idea of Lift. And I hope that’s sinking to some extent, we're going to discuss some other case studies. The idea that value is not really generated at C-level, consumer level. So capitalism is departing from that level because there's no interest anymore. And another case that we think it's relevant is Rakuten, which means optimism in Japanese. It's a great name for a company. And it started as something similar to Amazon back in the days. You might have heard of Rakuten, I guess.
It was already more lifted than Amazon when it started. It never ran its own warehouses or it never bought any inventory. It was operated as a platform, where independent merchants sell to customers, and then Rakuten would take a cut from those transactions. Since the 90s, Rakuten has acted more like a broker. It directs consumer traffic to third party merchants, and it earns commissions and other types of fees, part of which it rebates to consumers as loyalty points. We'll talk about these later again, the idea of loyalty points.
If a first order market is B2C, we discussed these earlier. Marketplace as a service is B2B, business to business in that the customers here are the businesses who sell the products through Rakuten itself. And here is where the critic from platform capitalism gets a little bit stuck. We sense in the discourse that a platform is all about consumers. That it generates a consumer facing entity that they engage in social engineering to trap consumers for network effect. And, for someone, for example, like Yanis Varoufakis, you may have heard about Technofeudalism, it's even more extreme.
For people talking about quote unquote Technofeudalism, as I was pointing out, these platforms are actually evolving into micro political environments. They are so invested in customers that they are literally intervening in their lives, vertically integrated into full stack providers and sustainers of human welfare. And this seems to us actually wrong. It's kind of a simplistic take on the matter, we believe. Rakuten gets paid precisely when customers leave the platform. Rakuten is providing a mesh like aggregator and organisational layer for other businesses.
Platforms to us are not vertically integrated walls, they're just extremely porous. Yes, we agree Rakuten does provide a platform, several of them actually, but it's by virtue of its own platformness, a complete world. Is that the case? I mean, is it a gesture towards a sort of ‘worldness’, if you want? We believe that's not the case. For us, a platform, and this is a very important idea in the book I think, is about being incomplete. It's about surface area for integration. A platform is not a matrix of relationalities.
A platform is a vibe. It's about partnerships, channels, alliances through which a given platform provides as many management solutions as it's receptive to. It's about fertility in a way. And if you go to Rakuten’s website, you'll see that this is exactly the case. It's a cascading wall of brand associations with which Rakuten is engaged in a circular flow of revenue. The more revenue moves, the more velocity it has, the more randomized its behavior, the more of itself is actually created. Let us consider now a different example.
I'm coming from Hong Kong. So this is something that is kind of relevant. We've both been to China a few years ago. And we're interested in what's going on there. And the example that we wanted to discuss after Rakuten is WeChat, which was originally launched only 14 years ago, first, as Weixin within the Chinese market. It seems to be one of the most consolidated platform environments on the planet. If you try to take the Technofeudalist approach to WeChat, you know, you would start pointing out to the containerization of 1.3 billion monthly active users into an object that interfaces with the Chinese state apparatus, if you want, as evidence for the idea that WeChat’s vision is not only to facilitate, but moderate the world space of contemporary China, of the contemporary Chinese public space.
But we believe that that's not what WeChat is. A little bit like Rakuten, WeChat is an application marketplace. I don’t know if you've ever used WeChat.
Marek:
It’s the craziest app of all times.
Roberto:
So WeChat combines into a single space everything that you have on your phones. You only need one app. But what we're trying to argue here is that WeChat kind of limits itself to a kind of identity provider, for which third party applications engage with users. They developed something that was called the mini programs, an environment that opened up the space WeChat. that opened up the space WeChat. that opened up the space WeChat. And it was created into a kind of general purpose app store through which WeChat could standardize protocols around information transfer and identity.
If you were taking Varoufakis’ approach, you would say that WeChat is actively invested in managing services like transportation, for example, vertically integrated itself into the physical lives of its users in order to maximally extract from their attention? But this is not what is actually happening. Instead, what WeChat is doing is, for example, with partners like Didi, which is a taxi driving company, it offers taxi services within an interface that looks identical to Didi's native application, but is actually served up within the raw interface of WeChat.
Luxury brands like Gucci and Fendi have their own mini programs for marketing and e-commerce, and there's 20 million businesses running...
Marek:
20 million businesses!
Roberto:
... official accounts...
Marek:
It’s like an app store with 20 million business, it’s crazy.
Roberto:
... running within Wechat. Far from tightening, if you want the Feudal grip, this mini program framework abstracts user interactions into a programmable layer open to outside developers. So WeChat in a way is more of a data broker. It's a hinge between the businesses that operate within it and the state. It's a hinge that capitalizes through the dual WeChat-Weixin structure. So WeChat was first released as Weixin within China. Then as WeChat for the international market on the complexity of the Great Firewall.
The two systems, one app design, monetizes governance and regulatory complexity through an indifferent approach. It’s an arbitrage on sovereignty itself and it lifts from the brittleness of an apparent multipolarity. Now, WeChat is owned by Tencent, which is a huge company based in Shenzhen. And when you look at Tencent’s revenue streams, it all becomes even more clear. Tencent does not make money by rallying fleets of, you know, taxis or delivering groceries. It makes money by taking a cut from online games, subscriptions and premium services run by other people.
The other large pillar is financial services, which is WeChat Pay. And this is something that hosts trillions of dollars annually, clipping a fraction from every transaction. In 2023, Tencent's fintech and business services segment, dominated by WeChat Pay as we just discussed, generated 203 billion RMB, becoming a financial gatekeeper and surpassing its traditionally large gaming segments’ revenue. In both cases, thinking about the gaming kind of side of it and the financial side of it, the logic is the same.
Tencent charges businesses for how they interact with people. So what looks like total consolidation from the outside is in fact Lift. WeChat is not pulling people into a closed platform ecosystem. It is rather abstracting their interactions into a programmable data layer, one that sits above production and consumption. WeChat in a way abstracts human activities into digital data flows that Tencent can intermediate.
Marek:
Let's look at Nvidia, which is one of the most crazy and interesting companies out there. That's obviously the hardware company that’s behind, quote unquote, behind the AI boom, especially the US, but obviously everywhere. So what does Nvidia make? GPUs, the processing units that are required for hardware to do AI, essentially. But what's interesting is that Nvidia obviously doesn't make those chips. They sell Taiwanese chips made by TSMC. And while they license designs to TSMC, the actual thing that Nvidia does, the value that they ultimately provide is literally software.
It's called Cuda. It's this software framework that transformed these chips coming from Taiwan into the brains behind generative AI. And so in 2022, Nvidia announced that it's going to mature in what they call managed services, a.k.a GPU as a service. You know, spinning up all these different platforms where you can basically, you know, rent their attention and generate inference using them. So a given kind of business out there could buy a license in order to deploy AI on remotely managed Nvidia hardware, essentially.
So it's a live transition. You're moving away from a dollar per GPU relationship to a $1 per GPU hour relationship. That’s Lift again, you're moving from paying for a thing once to paying for access to a thing over time. But even this is too direct, like Nvidia doesn't even want to do this, instead, like an insane, unthinkably enormous 88% of Nvidia's revenue comes from their quote-unquote data center segment, which would make a lot of sense if Nvidia actually owned and operated data centers. This is like the most valuable company in the US.
You know, this is like literally the thing keeping the American economy afloat right now. They literally make 88% of their revenue from two companies for their data centers usage. And they don't maintain, they don’t operate the data centers. They basically just license them out. So they provide licensing, certification and software management on top of data centers owned by other people, third party co-location partners who then lease and disperse the compute back to Nvidia for a fee. There's a lot of people out there, especially in like continental Europe, we hear the kind of critique of this idea that there's this thing called Technocapitalism, which is this hovering black steel war making titan with Elon Musk ketted out at the helm, who is powering it based on the desire to make intentionally, you know, bad, cruel decisions against human individual liberty.
Like that's the perception I get when talking about Technocapitalism, pretty much anywhere but kind of especially here. I love to point to, like the most important company for the American economy right now, which is Nvidia and its absolute total disregard, in effect, total non-engagement with the American consumer class, let alone any kind of psychosocial ideological manipulation of that consumer class. If we want to do something, like if we think that the current situation isn't good, which maybe we kind of agree is the case, we need to be able to extend our critical capabilities beyond our own experience of capitalism, our subjective experiences of capitalism, because we unfortunately aren't at the center of these things.
We're actually almost completely irrelevant to the way that these things make money. It's not about us. And that's really the problem.
Roberto:
So before we get into some really lifted stuff, we wanted to know that any business can lift at any scale and that this is an inbuilt compulsion of businesses itself, according to a theory of Exocapitalism. So we're going to go into something else. I might seem a bit, maybe, you know... it might seem unrelated. And that's pornography. Pornography in industry. And you can see some really classic lift dynamics here. For example, you saw individual performers or creators associated with this or that representation engage with studios who produced and sold physical merchandise.
And famously, Girls Gone Wild were sued by the American Justice Department for operating a pseudo continuity program where you might have thought you bought a single DVD, but instead you were automatically signing up for a membership with auto renew but instead you were automatically signing up for a membership with auto renew and therefore purchasing a recurring revenue resource. As pornography became increasingly digitized, most mainstream producers stuck to subscription based models, leveraging platforms like Pornhub as freemium marketing spaces, where clips could be shared to draw traffic towards their own studios.
Sole proprietors, in their part, could leverage that kind of creator mode to collect some drippings from the ad revenue the videos generated on the platform. In this sense, pornography producers were already lifting from B2C style engagement, instead selling their videos essentially to advertisers. But as pornography transitioned into more exclusive content, it lifted. In fact, the ’exclusive’ in ‘exclusive content’ precisely names a transition from ownership to contractualized leased access to pornography as a service.
As the popularity, for example, of OnlyFans exploded, especially during the Covid period, many creators found themselves running viable small businesses, discovering that the real money was not in the first party production of pornographic material, but in dedicated OnlyFans management agencies for the heavy lifting of businesses and marketing tasks. In turn, a meta management business model has been cascading upwards, In turn, a meta management business model has been cascading upwards, coaching OnlyFans managers on how to create and sustain an OnlyFans management business.
Marek:
This is the most depressing and upsetting video I have ever seen in my life, if you want to ruin your evening. It's OnlyFans management and management coaching. So it's a type of management coaching for people who are going to manage other people who are building OnlyFans management companies. And it's led by Markuss Hussle. And it's like the most misogynistic, evil content in the whole universe, but, that's like the stack. That's what gets built from all of this, it's the craziest thing. Let’s talk about Starbucks.
We're reaching a kind of high point here. We're reaching a kind of high point here. So Starbucks is super lifted. It forces customers off. And if you want to use the app, you upload a block of money at a time, right? It's usually 25 USD, in exchange for loyalty points, which are called stars, which is really cool because those stars are literally bucks in a way. Like it's really beautiful, kind of poetic. So basically, customers are basically lending money to Starbucks, $25 at a time in advance of receiving any coffee.
As of earlier this year, Starbucks was sitting on 2 billion USD in-store value balances from customers. That's interest free capital, baby. You can do whatever you want with that. You can redeem that cash. You can use it for operations. You can reinvest it until customers ultimately redeem it. And unlike a traditional bank, you don't pay any interest or anything like that. They don't have to do anything. They just chill with your money. Usually there's about 200 million plus a breakage a year. Starbucks is also really structurally lifted through licensing, so 48% of Starbucks stores aren't actually managed by Starbucks, but are actually licensed to third parties.
And so those licensed stores are generating reward points, royalties, margin on Starbucks supplied product. But they keep all the kind of operating costs, all that stuff with the actual partner itself. So Starbucks is monetizing brand, management software, recipes, you know, IP equipment and raw supply at the meta layer, while all the other people are doing the hard work of actually dealing with customers and making coffee. So, is Starbucks a coffee company? It's a question I want to ask. There’s coffee involved for sure, though you'd probably disagree that it's coffee, but it's definitely connected to revenue in some way, but is it determinative of revenue?
It's kind of more like a software company in a lot of respects, or it's a lot more like a bank. And in these capacities, yes, there’s labor involved, but who's doing it? Because if it's a bank, is it the barista labor for the bank? No. It's the kind of banking labor that materially matters to Starbucks, in the last instance. So the question is, what kind of labor is important? And if we're people in this room who care about labor, it’s a real question, what kind of labor power is important in this dynamic?
So when we talk about the supply chain, which is, you know, coffee producers shipping the coffee, blah, blah, blah, all this kind of stuff, it's really easy to kind of think of it as this long vertical link, but it isn't. It's like a deeply, you know, interlinked nested array of lifted things, each of which are supporting a lot of multiple redundancies in terms of their downstream upstream dependencies. So McKenzie Wark has a concept called Vectorialism, which means that it's really, really hard to, her suggestion is that it's really hard to break the supply chain, because at the end of the day, if one supply chain part breaks, you're not working with them anyway.
You're working with like a procurement agency, like 17 orders higher, and they're going to find a partner who's going to move the supply chain elsewhere and find a cheaper, better option. So the contents of the supply, like the thing being supplied, doesn't actually matter in its specificity in any way. And they aren't distinct. They aren't individual in any way. They're actually fungible, interchangeable abstract vectors. That's what McKenzey work means by a vectorialisation. So when you’re sitting on a supply chain, you're not thinking about the things that are actually being supplied to you.
You're thinking about lines, basically. You're thinking about abstract motion. But this gets so much more intense when you understand through Lift, because not only does the produced good not matter in terms of its particularity, but the connection between produced good and revenue is unspooling. As you lift, you find yourself participating not in a commodities economy but an economy of stakes, alliances, speculation, all of which are not only removed from the commodities economy but are actually increasingly self-sufficient, and they capitalize purely on what Marx called fictitious capital, but is in fact the only real capital. Arbitrage.
Roberto:
Arbitrage.
Marek:
Arbitrage. These are like the two most insane examples.
Roberto:
So another example we wanted to discuss is Japan's SoftBank, which takes this movement to an even more extreme level. SoftBank began as a kind of software license manager and distributor, already a kind of weird and lifted relationship to software, which it managed by selling magazines with floppy disks in them. And this was back in the 90s, early 90s. Later on in the 90s, SoftBank shifted its focus dramatically. It managed to get a lot of funding and use that to finance speculative bets on other companies.
And the idea here was to arbitrage between the market difference between the US and Japan. And that was a huge move. This kind of hit its peak with SoftBank’s legendary now bet on Alibaba. They bought 20 million USD on a stake in 2000. By the year 2014 that have become 50 billion. Other killer acquisitions that they made of international NMCs, a state against Japanese valuations include Vodafone Japan turning SoftBank for a while into a management layer on top of a management layer for telecommunications.
And it descends; SoftBank escaped operating as a platform or a major services organization altogether, taking the leap directly into highly informed strategic investments and affiliations, strictly capitalizing on raw geopolitics itself. Things got far more lifted in the 2010s, with a huge amount of external funding that was coming from Saudi Arabia’s PIF and Abu Dhabi’s Mubadala And then SoftBank actually became an investment fund. They use all of this cash to bet on companies like Uber, WeWork, DoorDash, Coupang finally acquiring pieces of Aam and now ultimately joining Oracle and OpenAI in the 500 billion Stargate project, a funding scheme that seeks to capitalize on the extraordinary volatility of the international AI boom.
But it was their mistake to rate SoftBank's maturation from software to software rights management to speculative capital management as a kind of late 20th century phenomenon. Almost every hallmark law of conventional industrial capitalism went through the same process. That's the argument here. From Berkshire Hathaway’s mid century transition, from textile, mining to insurance powered capital flow allocator to ambiguous multinational holding company to, you know, companies like Hudson in early 20th century transition from real estate management.
And we can think of a late 19th century maturation of US Standard Oil for refining and operating and, eventually became divided into what is today Exxon and Chevron, who operate in a similar way. But we could also think of the 15th and 16th century Medici state’s transition from cloth to early financial services. In each of these cases, the underlying good became less and less relevant to the underlying social value. Given these proportionally limiting returns, scaling instead through the abstraction of the product, the reinvestment of capital into itself, and the establishment of reciprocal relationships within the broader plane of business as such.
Marek:
This one is so sick, this is Citadel, they are giant a hedge fund, who used to be like a bank, but now they're just a fully automatic money machine. So Citadel Securities, they use advanced algorithms to trade stocks, options and other assets at crazy speeds. Usually they hold a stock for like a couple seconds at hard max. And the logic behind these trades is not about the long term prospects of a company, that are like investing in a company per se, or any kind of underlying economic factors at all.
They're literally just trying to exploit tiny discrepancies in prices. They're maybe looking at patterns in orders and using that to take bets or other kinds of statistical opportunities. So the idea here is that Citadel is using randomness itself as a source of profit. And at the center of high frequency trading is the concept of arbitrage. So if the price of gold in New York is momentarily lower than the price of gold in London, you can use an algorithm right to buy in one market and sell it at the same time, and just immediately make money and just kind of keep doing that until the market ultimately corrects itself.
But that's still connected to a real thing, which is gold. And Citadel is not. It’s way more abstract than that, usually based on purely statistical grounds. So the easiest way you can think about it is like maybe they're looking at one asset and they notice that whenever this asset goes up, this asset goes down. So maybe they have an algorithm that's looking at this asset that goes down by just a little bit. We'll place a bet on this one and then we'll trade off. So that's one way to do it. Another way you could do it is you could train a machine learning model, for example, on a ton of market data and then, you know, listen to that model, if it tells you to buy something, you buy it.
But even that's too predictable in that those fluctuations in value are too dependent on real world conditions, because other people might be doing that too. So if you have a bunch of people looking at two connected assets, one person might, you know, you might take a bet on that, another person might take a bet on that, and then it'll create this kind of... It will smooth out that randomness and you won't be able to profit from that at all. So, there’s this crazy euphemism in finance called a dark pool, which is like the most beautiful way of talking in finance ever, where you actually place a whole bunch of bets together, in a kind of very secret place, and you just exercise them, and then you get the fuck out, that's just how they do it, because they don't want anyone else to be aware that those transactions are happening, because that will disturb the market and they will be able to predict that and capitalize on it et cetera.
They're actually really hiding these things in kind of secret ways. But that's not even the coolest way. The best way to do it is just to use math. Like just use stochastics to look at a random number line and make simple inference based on the inherent patterns in randomness itself. I don't know how deep you guys are into math, but randomness is cool because randomness is not random. It's actually really almost impossible to make a truly kind of random number. Randomness has all kinds of internal turbulences, which we understand through a branch of mathematics called stochastics.
So Citadel doesn't care what the underlying asset that they're betting on is. They're just looking at a number line, like a random curve. They might not even know what the asset is. They might not even know what they're investing in for a couple seconds. The content of the trades is irrelevant. They're generating profit because the market's continuous fluctuations provide lots of opportunities to just get in, get out based on just straight up prediction methods. And so here we're really arriving at a kind of circular flow of capital in which it's just movement alone that's generating capital.
And what's cool or what's crazy is that you can profit under any market conditions. You might think like number go up, everyone's making a profit. Number go down, nobody's making a profit. But that's not how it works. It's like you could bet against things, right? It's really easy to place a put position, short something, make a bet basically that the price of something is going to go down. If you make a bet that the price is going to go down and it goes down, you make money, right? So what's crazy is that even if the market is suffering, even if the big bubble just popped, a bunch of people might have just made a shit ton of money, right?
So it really doesn't matter how the market moves. It matters that the market moves. Everything underneath it, the machines, the human sweat, the blood stolen through colonial violence, the service workers at pharmacies or kiosks, the software engineers, the H.R. managers, the humming GPUs enhanced by extracted neodymium, all of it bubbles up to the level of a random number generator, against which bets can be made simply through statistical inference. None of it matters, all of it is fungible, and you don't need a computer to do this.
You can do capitalism on anything. You can do it at a horse race. You could do it by holding up cards in a crowded cigar smoking room. You could do it by holding up cards in a crowded cigar smoking room. You could do it by moving seashells around the west coast of Africa. It literally doesn't matter. The substrate literally doesn't matter. And in our liberation from the material surface, we encountered a layering process that can elastically compound upwards forever.
Roberto:
You may argue that most of these examples are too recent and that we're just looking at something that is a recent phenomenon, and we are focusing on that and we're trying to explain it somehow. We're trying to work against that idea. So we are trying to define all the history of what capitalism may be as a form of software economy. We are trying to look at abstraction as a kind of driving force that is not new. Derivatives of financial capitalism are just one more form of engaging with that sort of reality that underpins as Arbitrage, the difference between values, the kind of translatability of an entity between different worlds, different markets, different margins, etc.
And so at the end, this kind of final example that we want to mention, briefly, it's from the 18th century, which is the South Sea Company. And we argue that it practiced an extreme form of arbitrage with no reliance on computation or silicon or the internet. The South Sea Company began as a way to repackage British government debit by mobilizing permissions obtained via the 1713 Treaty of Utrecht around the sale of enslaved African people to Spanish colonies, which is kind of dark, fucked up. The British government had incurred a lot of debt during the war of Spanish Succession, which ended with the Treaty of Utrecht.
And in doing so, it had issued an extraordinary volume of IOUs to the providers it owed money to. It proposed a relationship with the the South Sea Company, where anyone with an IOU to the government could exchange that IOU for equity in the transatlantic slave trade, which at the time was only starting to actually operate at scale. It was quite small. This was an almost completely speculative, hype driven investment, and this is sad to say, because of the kind of underpinning reality of what they were betting on.
All of a sudden, the South Sea Company became enormous, despite an almost complete hold on the shipping operation. And not only that, they could reference the equity they had raised in order to sell new shares to pull in even more money. Essentially raising the financing volatility of their own stock. When this resulted in a financial disaster at the end, as most bubbles do, that did not compromise the underpinning dynamics that were in play. Nor does it mean that a lot of people didn't make a financial killing out of exchange.
And you might have seen the meme of Newton, who was involved in this.
Marek:
This is not how to gamble on the stock market by the way. This is the worst way to do it.
Roberto:
So, as Marek was pointing out, the key element here is that arbitrage can work regardless of whether the line goes up or if it goes down. The only important thing is the movement of the line. It just has to move. And this is where Exocapitalism starts as a theoretical project, with arbitrage as the kernel of capitalism, with the idea that corporate interests tend toward the liquid, the recurrent, the independent, the vectorial, the infinite, and in doing so it tends away from the soil, the land, the blood and the sweat.
This is not to suggest that capitalism is good. Not at all. We are not claiming that. I hope that that's clear. But absolutely it does suggest that capitalism is, that it exists as a thing on its own terms.
Marek:
So finishing this up now, it's a problematic book, obviously. It does a lot of things that you're kind of not allowed to do anymore, things you're not really allowed to say, I guess, we're committing a lot of sins against Marx, who you could say is the book's enemy in a lot of respects, but is actually someone we have a lot of respect for. I think our enemy instead is the calcification of leftist thought strictly against this kind of social critique of capitalism, the idea that, you know, we're people, capitalism is a socially manifested thing, and this entire critique is built on an understanding of what capitalism is to the exchange of commodities, and it becomes a kind of evangelical literalism of exegetical, biblical text, Das Kapital.
And this calcification is so extreme and dogmatic that I find myself struggling to commit to the left, to be honest, especially to the extent that the left seems to kind of dither in superficial moral conflict between absolute good and bad, big and small, innocent and corruptive, victim and perpetrator and the bad, the corruptive, the perpetrator, they tend to collapse into a kind of singularity of evil. And whereas I'm describable through so many of the vectors of evil, so it's a natural inclination for me to want to play defense, but at the same time, this book is not an argument against vilification, but rather for a true polytheistic pantheon of evil, one where different aspects of our enemy can be understood through avatars with different motivations, different abilities, and different dispositions.
And in some respects, maybe ironically, we're following from Kimberlé Crenshaw's definition of intersectionality, which can be refracted into the idea that there is this great difference within the spectrum of oppression and oppressors, and that this difference may aggregate around powerful social contractors.
Roberto:
But this book is not about oppression, which I think is exactly what makes it a kind of unique contribution to the discourse. It is instead a thought experiment about the internal mechanisms of capitalism. There are thousands of books you can read about the oppressive externalities of capitalism. We can recommend a few. I'm sure that you probably know most of them. But if we understand capitalism strictly through its capacity to oppress, we argue that we limit ourselves to an essentially reactionary position.
No, if capitalism is your enemy, you should know your enemy. And you should understand your enemy not only through its tactical advantages on the battlefield, nor through the recounting of thousands of lost battles, but through an honest accounting of its motivations, at its own terms. We suggest that our enemy is more likely ourselves. But this is another long lecture for a different day.
Marek:
Thank you.
©2026, Berlin/Nicosia
hello@becoming.press
@becoming.press
Prolegomena
Impressum
Submissions
Distribution
Imprints
Agenda
Contact
Newsletter
Subscription
Events
Radio • On-Demand
ÊT/RE
Materials
Immaterials
Seminars
Cinematics
Nodes
Search