12.11.2026

Grappling with digital speed

Yassamine Mather dipped into the sessions on imperialism and Iran. They were not only poorly presented, but the politics were awful. So, instead, she took to the ‘AI and capitalism’ stream. It proved much more interesting

The 23rd annual Historical Materialism conference took place at SOAS University in London from November 5 to 8, under the title ‘Planetary warscapes ablaze: imperial mutations in infernal times’. It was the largest conference to date, with 317 panels and approximately 1,600 participants.

Imperialism and war were central themes. The ‘Call for papers’ hub presented US hegemonic decline as producing sharper inter-imperial rivalry rather than a more peaceful multipolar order. It highlighted Gaza, the US-Israeli war on Iran, Ukraine, tensions over Greenland and the confrontation with China. A dedicated ‘Imperialism and anti-imperialism’ stream addressed these developments, including debates over whether China should be considered imperialist and how international solidarity might be built without falling into liberal humanitarianism or uncritical campism.

However, I will concentrate on the sessions I attended. I had intended to follow the series on Iran and the war, but the second session was so appalling in both content and presentation that I decided to follow the ‘AI and capitalism’ steam instead.

This examined artificial intelligence in relation to capitalist accumulation, corporate power, labour, social reproduction, imperialism and political resistance. A central argument running through the discussion was that AI cannot be understood independently of the social relations within which it develops. The enormous investment in computing infrastructure, the growing influence of a small number of technology corporations and the integration of algorithmic systems into workplaces, households and state institutions reflect the priorities of capitalist accumulation, rather than technological innovation alone.

However, important theoretical differences emerged. Does AI represent a qualitatively new stage of capitalism? Is the concentration of digital power evidence of a transition towards techno-feudalism, or does the extraction of rent remain subordinate to the production and appropriation of surplus value? Can AI restore profitability and initiate a new period of capitalist expansion, or is the extraordinary investment in the sector creating another speculative bubble?

Three discussions provided a useful framework for examining these questions: the political economy of AI and corporate power; the debate over techno-feudalism and capitalist profitability; and the wider analysis of algorithmic capital, social reproduction and resistance.

Corporate power

The session on ‘The rulers: corporate power in the age of AI and the cloud’ featured Cecilia Rikap and Jodi Dean. The session provided a starting point for examining the concentration of economic and political power in contemporary digital capitalism.

The central issue is not simply whether large technology corporations possess excessive market power, but how their control over cloud computing, digital infrastructure, data and AI systems enables them to shape the conditions under which other capitals operate.

In her introduction, comrade Rikap argued that developing AI requires enormous computing capacity, specialised hardware, data centres and investment. These requirements create high barriers to entry. Although AI is frequently promoted as a means of empowering small businesses, its underlying infrastructure is concentrated in the hands of a few corporations.

Companies controlling cloud infrastructure and computing resources can determine the terms on which other businesses access essential technological services. They do more than sell products: they influence the environment in which production, communication and scientific research take place.

From a Marxist political economy perspective, this demonstrates the continuing tendency towards the concentration and centralisation of capital. Competition has not disappeared, but increasingly takes place within structures defined by unequal control over data, infrastructure and finance.

Cecilia Rikap challenged the argument that the dominance of big tech indicates a transition to an entirely new economic system. Economic concentration and centralisation are longstanding features of capitalism. Powerful technology monopolies therefore represent an intensification of existing capitalist tendencies rather than necessarily constituting a break with the system.

Another important issue concerns the relationship between corporate technological power and the state. AI development depends heavily on state investment, publicly funded research, military expenditure and geopolitical competition. According to comrade Rikap, this relationship is reciprocal. Governments increasingly rely on privately controlled digital infrastructure, while technology companies benefit from public contracts, favourable regulation and military procurement. The integration of AI into intelligence, surveillance, military planning and weapons systems demonstrates the political significance of this alliance.

For countries outside the principal centres of technological accumulation, these developments deepen digital dependency. Access to advanced computing, proprietary models and cloud services remains largely controlled by corporations based in dominant capitalist economies, reinforcing unequal international divisions of labour and technology. Rikap added that regulation may limit particular corporate abuses without transforming ownership of the infrastructure itself.

Jodi Dean, author of Capital’s grave (Verso, 2025), challenged comrade Rikap’s concept of “intellectual monopoly capitalism”, arguing instead that contemporary capitalism is moving towards neo-feudalism.

I found her intervention on political strategy more persuasive. She emphasised the importance of party organisation, whereas comrade Rikap advocated digital sovereignty as a means of democratising technological governance without embracing techno-nationalism.

The differences between these approaches raise questions extending beyond the classification of contemporary capitalism: what political forces could challenge the concentration of technological power, and through what forms of organisation?

Techno-feudalism

In a related session, ‘Techno-feudalism: rent or profits?’, the panel included Michael Roberts, Renu Gehring, Tavo Espinosa and AK Norris. Comrade Roberts questioned whether the growing power of big tech signifies a movement beyond capitalism.

According to the techno-feudalism thesis, digital platforms and technology monopolies increasingly resemble landlords controlling digital territory. Businesses and individuals must pay for access to the infrastructure, markets and services these corporations dominate. Rent, rather than profit derived from capitalist production, supposedly becomes the principal organising mechanism.

This interpretation draws attention to the revenues platform owners extract through intellectual property rights, proprietary software, digital marketplaces and cloud infrastructure. The question, however, is whether these forms of monopoly appropriation amount to a new mode of production. Comrade Roberts argued that they do not.

Capitalism remains organised around the production and appropriation of surplus value, even when the mechanisms through which value is distributed become increasingly complex. Rent and monopoly revenue have always existed alongside industrial production, competition and wage labour. Their expansion does not, in itself, demonstrate that capitalist relations have been replaced by feudal ones.

This distinction is particularly important in discussions of AI, which is sometimes presented as though it creates wealth independently of material production and human labour.

In reality, AI depends on extensive physical infrastructure: data centres requiring construction and maintenance; semiconductors produced through complex industrial processes; electricity, water and raw materials; and workers engaged in software development, logistics, content moderation and data labelling. Knowledge and algorithms acquire economic significance through their incorporation into capitalist production and their contribution to profitability.

A further question concerns whether the extraordinary investment in AI will generate sufficient returns. Companies anticipate that automation, productivity improvements and new markets will increase profits. Yet gains for individual firms, achieved through lower labour costs or a greater market share, do not necessarily translate into higher profitability across the economy.

In Marxist economics, replacing living labour with machinery can increase productivity, while raising the proportion of capital invested in machinery relative to labour. The implications for the general rate of profit remain contested. This contradiction is particularly relevant to AI, where enormous expenditure on computing infrastructure is justified by expectations of future revenue and productivity gains that remain uncertain. The possibility of an AI investment bubble cannot therefore be dismissed. Corporate valuations and infrastructure spending may be growing faster than the sector’s capacity to generate sustainable profits.

Nevertheless, a collapse in speculative valuations would not render AI economically irrelevant. Previous technological transformations have also been accompanied by speculative booms. Financial crashes can destroy inflated paper values, while leaving behind infrastructure and technologies subsequently incorporated into capitalist production. The distinction between financial speculation and longer-term technological change is essential.

The debate also has an international dimension. Corporations and states controlling digital infrastructure may appropriate revenues generated through productive activity elsewhere. Understanding this process requires distinguishing between productive labour, monopoly power, rent, profit and imperialism rather than treating every form of appropriation as equivalent.

Ultimately, the techno-feudalism debate is not simply a disagreement over terminology. Different explanations suggest different political responses. If the dominance of technology corporations represents a system primarily based on digital rent, limiting monopoly power or restoring competition might appear to be the appropriate response. If, however, these developments remain rooted in capitalist relations of production, the fundamental questions concern ownership, exploitation, accumulation and class power.

Comrade Roberts’ argument was that AI transforms the exercise of capitalist power without abolishing capitalist relations of production.

Algorithmic capital

Jonathan Martineau presented ‘Algorithmic capital: accumulation, power and resistance in the age of AI’, co-authored with Jonathan Durand Folco.

He described algorithmic capital as a transformation affecting production, political power, social relations and humanity’s relationship with nature. The theoretical framework draws on Nancy Fraser’s understanding of capitalism as an institutionalised social order, encompassing not only markets and wage labour, but also the state, social reproduction and the natural conditions that make accumulation possible.

Building on this approach, the authors identify four interconnected dimensions of algorithmic capital: accumulation, power, social mediation and the relationship with nature. According to comrade Martineau, the first dimension concerns the incorporation of data extraction into capitalist accumulation. Digital platforms continuously collect information generated through work, consumption, communication and everyday activities, transforming it into commercial assets, predictive systems and targeted advertising. While recognising that theories of surveillance capitalism have drawn attention to behavioural data extraction, the authors criticise approaches that neglect the centrality of human labour. Data does not exist independently of social activity. Its production, annotation, organisation and application depend on workers ranging from highly skilled software engineers to poorly paid labourers engaged in content moderation and data labelling.

Rather than eliminating exploitation, platform labour and algorithmic management reorganise and intensify it. Algorithmic technologies also extend capitalist imperatives into activities traditionally located outside direct wage relations. Through household devices, digital platforms and online services, capital increasingly monitors and mediates care work, emotional relationships, domestic tasks and leisure. Capitalism has always depended on unpaid or underpaid labour, particularly in social reproduction. What is changing is capital’s capacity to collect information about these activities and turn aspects of everyday life into opportunities for commercial extraction. This development raises questions about the boundaries between production and reproduction, paid and unpaid labour, and the growing commodification of social relations.

Algorithms increasingly organise decision-making across workplaces, consumer markets and state institutions. Under algorithmic management, decisions concerning wages, schedules, performance assessments and dismissals may appear automated or technically determined.

However, according to comrade Martineau, treating algorithms as autonomous forces creates a misleading impression of neutrality. These systems are designed, implemented and controlled within existing structures of managerial and political authority. Their decisions remain political and can therefore be challenged through collective organisation and worker resistance.

Algorithmic power also extends beyond the workplace. Digital systems influence cultural consumption, communication and self-perception, while their incorporation into surveillance and military technologies strengthens the relationship between corporate profitability and state coercion. The authors argue that privately owned platforms do not merely transmit social relations: they increasingly structure the conditions under which those relations develop.

Technologies embody choices shaped by the economic and political priorities of those who finance, design and deploy them. Nevertheless, a Marxist analysis must avoid technological determinism: algorithms are not independent actors governing society.

The supposedly immaterial digital economy also depends on extensive physical infrastructure, including data centres, electricity networks, water supplies, semiconductor manufacturing and mineral extraction. Comrade Martineau also spoke about AI expansion’s significant environmental consequences.

These burdens are distributed unequally. Under conditions of digital imperialism, resource extraction and labour-intensive activities are frequently concentrated in peripheral economies, while ownership, profits and technological control remain disproportionately located in dominant capitalist centres. The author then addressed possibilities for resistance.

The objective is not necessarily to reject technology, but to challenge the social relations determining its development and use. Struggles surrounding data centres, surveillance and algorithmic management, including those in Chile and Ireland, demonstrate how technological questions intersect with labour organisation, environmental politics and community resistance.

Three strategies emerged from this discussion. The first was technical sobriety: questioning whether every technically possible innovation should be developed, and prioritising social needs over profit-driven technological expansion. The second was public ownership and socialisation. Regulation and taxation may constrain corporate behaviour, but they do not necessarily transform private control over digital infrastructure and data. The third was coalition-building between technology workers, platform drivers, logistics workers, communities affected by extraction and movements opposing surveillance. Such alliances could connect struggles that are often treated separately, despite their common relationship to capitalist accumulation.

These discussions provided a valuable analysis of AI’s place within contemporary capitalism. Yet they also left me with a question that I believe the left has not addressed adequately: how can we use existing AI technologies to strengthen socialist organisation and political struggle?

The left

I did learn a lot from the sessions I attended; however, I must admit that I question aspects of the left’s approach to artificial intelligence. Much of what I read and hear is dominated by fear: fear of unemployment, surveillance, corporate control, ideological manipulation and the enormous power accumulated by a handful of technology companies. These concerns are entirely legitimate. However, there is remarkably little discussion of how the left might use the same technology to its own advantage, including in the struggle against capitalism itself.

In response to the concentration of AI technology in corporate hands, we are frequently offered familiar solutions centred on regulation, public ownership and democratic accountability.

Cecilia Rikap and others, for example, advocate democratic control over the infrastructure underlying large language models (LLMs). But what exactly would democratic ownership mean in this context? Who would exercise control: national governments, international institutions, workers’ organisations or some new form of transnational democratic authority? Given the global infrastructure on which AI depends, the question is far from straightforward.

Michael Roberts has proposed a more radical alternative: bringing the so-called ‘Magnificent Seven’ technology companies - Apple, Microsoft, Alphabet (Google), Amazon, Meta, Nvidia and Tesla - into public ownership under democratic workers’ and public control, potentially through expropriation without compensation to major shareholders. This is certainly a proposal worth supporting. However, its realisation would require a fundamental transformation of existing property relations and, ultimately, the overthrow of capitalism.

The question remains: what can the left do now, under existing capitalist conditions, rather than waiting for a future socialist transformation? I would argue that there is already considerable scope for using existing LLMs, including technologies developed by the very corporations we oppose, to strengthen anti-capitalist organisation, research, education and political intervention.

Part of the difficulty is that many social scientists writing about AI do not have a sufficiently detailed understanding of how the technology works, how rapidly it is changing or what can be built using it. This is not to dismiss political economy, sociology or historical analysis. These are essential to understanding the social consequences of AI. But they cannot substitute for technical knowledge of the systems being analysed. There is also the problem of time.

Books and academic articles necessarily take time to research, write and publish. In most areas of political theory, a year is a relatively short period. In AI development, however, it can encompass several generations of technological change. An analysis that was accurate when written may already be describing systems whose capabilities and limitations have substantially changed.

This does not make theoretical analysis obsolete. On the contrary, it makes closer collaboration between critical social scientists and technically knowledgeable software developers, computer scientists and AI researchers increasingly important.

There are already leftwing technical specialists who understand the possibilities of these systems and can imagine using them to challenge capitalist power. However, such individuals are often dispersed, work in isolation and are rarely taken seriously in broader political discussions. The left needs to bring these different forms of knowledge together rather than treating technical expertise and political theory as separate spheres.