Developments from August 20–24, 2026: the socialization of production and risk versus the concentration of ownership and power
Discussions of artificial intelligence usually focus on models: Which model is more powerful, and how much human work can it automate? Yet the developments of August 20–24 raise a more important question: Who controls the capacity for intelligent production? AI is no longer merely a software industry. Behind every advanced model stands a vast network of human knowledge, data, chips, electricity, land, water, data centers, financial capital, and labor. AI production is becoming increasingly social, while control over the conditions that turn these social resources into economic power can remain concentrated in the hands of a small number of corporations and owners of capital. This is the contradiction that assumes fundamental importance in the book A New Social Order in the Age of Consciousness.
From Owning Models to Owning Capacity
During the first stage of generative AI development, competition was organized largely around models. It is now clearer that possessing an advanced model is not enough. Large-scale operation requires computing capacity, and that capacity is produced through a combination of chips, energy, land, the electrical grid, data centers, and capital. The economic power of AI is therefore shifting from “ownership of the model” toward ownership of the material conditions required to run it. Even when knowledge and model weights are publicly available, openness does not necessarily distribute productive power if only a few companies can secure hundreds of thousands of GPUs, several gigawatts of electricity, and enormous data centers.
The Financialization of Intelligent-Production Capacity
The large-scale entry of finance into AI infrastructure is one of the clearest signs of this shift. NVIDIA, together with major financial institutions, has pursued mechanisms designed to mobilize more than $500 billion in third-party capital for AI infrastructure. What matters is not only the amount of capital. Compute is becoming an investable asset class. Computing capacity can be incorporated into long-term contracts, collateral, debt, and financial claims. A new chain is taking shape: infrastructure → computing capacity → future revenue → contracts and collateral → financial assets. This is the financialization of intelligent-production capacity.
Debt and Ownership Claims on Future Income
Debt issuance associated with AI development in the United States has reached roughly $220 billion in 2026, compared with about $12.5 billion in the previous year. AI infrastructure is no longer financed only through the retained earnings of technology companies. Bond markets, banks, investment funds, and private credit have entered the structure. An investor may not physically own a GPU or a data center, yet through a contract or financial instrument can acquire a claim on the future income generated by that capacity. Ownership thus expands from physical possession to contractual ownership and financial rights over future revenue.
From Chips to Land and Energy
NVIDIA’s expansion from selling chips into infrastructure is another sign of this trend. A GPU producer no longer has to confine itself to selling the means of production; it can participate in chips, financing, land, energy, and data centers. This is vertical concentration of ownership: instead of controlling only one link, a corporation gains a presence across several critical parts of the chain.
The Price of Compute and the Cost of Entering Production
Reports on August 22 indicated that the prices of some next-generation servers equipped with AI chips could rise by more than 15 percent, although the increase had not yet been officially confirmed at the time. If compute is becoming one of the essential means of production in the age of consciousness, the cost of accessing it becomes part of the cost of entering advanced production. A university, cooperative, small firm, or independent researcher may have access to papers, open-source code, and even model weights, but without sufficient chips, electricity, and capacity cannot compete on equal terms with owners of multibillion-dollar infrastructure. Knowledge may become more open while the material ability to use it effectively becomes more concentrated.
The Rise of Full-Stack Ownership
Developments on August 23 made this trend clearer. Alibaba announced plans to raise approximately $10.2 billion and direct the proceeds toward full-stack AI development, including chips, computing infrastructure, models, and AI deployment. The structural advantage of the future may belong to the company that controls the chain of chip → data center → cloud → model → platform → user → revenue. At this stage, ownership of a model becomes ownership of an intelligent-production ecosystem.
Data: A Product of Social Activity and Raw Material for AI
Developments on August 24 again highlighted the question of data ownership. Large organizational datasets—including communications, documents, and operational information—are acquiring economic value for the development of AI products. The fundamental question is who owns data produced through collective human activity and how the new value derived from it is distributed. In their daily activity, workers produce emails, files, solutions, and organizational experience; that activity becomes accumulated knowledge. If this knowledge later serves as raw material for AI, data cannot be treated merely as “fuel for AI.” Data itself is a product of human social activity.
Automation Is Not Only About Eliminating Jobs
AI is entering programming, technology services, logistics, and robotics at increasing speed. The drive to produce more with lower costs and fewer workers raises the fundamental question of productivity. If AI can produce the same output with less human labor time, productivity gains could reduce working hours, increase free time, and improve well-being. They could instead lead to workforce reductions, pressure on wages, and higher returns to capital. Technology does not determine the path by itself; ownership relations and power do.
Private Profits, Socialized Risk
When hundreds of billions of dollars in debt, investment funds, stock markets, banks, data centers, electrical grids, and real estate become tied to the AI investment cycle, its risks no longer remain private. If investment succeeds, a large share of the profit can remain private. In a crisis, however, the effects can spread to financial markets, pension funds, employment, energy prices, and the broader economy. A new contradiction emerges: the private concentration of profit versus the socialization of risk.
Two Currents Moving in Opposite Directions
The developments of these five days can best be understood as two currents. In the first, collective knowledge and research, data and social culture, human labor and software, together with energy, land, water, and public infrastructure, flow toward intelligent production: chips, data centers, computing capacity, and models. In the second, AI products and services are transformed into markets and revenue, contracts and collateral, debt and financial claims, capital accumulation, and concentrated ownership and power. At the same time, electricity and water consumption, grid costs, labor-market disruption, financial risk, and environmental consequences are distributed largely across society.
From Model Ownership to Ownership of the Conditions of Production
“Model ownership” is no longer sufficient to explain the political economy of AI. The deeper issue is ownership of the conditions that turn social knowledge into products and economic power. Knowledge, open-source software, and open-weight models may be widely accessible; but if turning them into economic power requires passage through GPUs, cloud systems, energy, data centers, capital, and platforms, those who own these conditions will wield power far beyond their apparent contribution to the production of knowledge. The central question is no longer only who owns artificial intelligence. It is: Who owns the capacity to transform social knowledge into economic power?
Connection to A New Social Order in the Age of Consciousness
If cognition, knowledge, and data become the most important productive forces of the new era, the historical contradiction becomes more visible. The intellectual raw materials of this system—language, science, culture, code, experience, and data—have been created across generations by millions and billions of people. Its material infrastructure also depends on human labor, universities, energy, land, water, the power grid, and society’s accumulated capital. Yet the conditions that transform this social whole into an economic product can remain in the hands of a small number of firms. Production, knowledge, data, costs, and risks become increasingly social, while ownership of capacity, income, and decision-making power can become increasingly private, financialized, and concentrated.
The Defining Question of the Age of Consciousness
Artificial intelligence creates extraordinary productive possibilities, but technological progress alone does not determine how these possibilities will be distributed. AI could reduce the amount of necessary human labor, expand education and health care, and make access to knowledge nearly universal. The same technology could also concentrate wealth, weaken workers’ bargaining power, and increase economic control. The difference between these paths will not be determined simply by model quality. It will be determined by the structure of ownership and social participation in power. If intelligence and knowledge are more than ever products of society, why should the capacity to turn them into economic power remain under the control of a small number of owners of capital? This is not a question of opposing technology. It is a question of how the immense productive forces unleashed by AI can be placed in the service of human flourishing. This is where the discussion of technology becomes a discussion of A New Social Order in the Age of Consciousness.
Leave a comment