Hamid Akhavi | September 24, 2026
Comprehensive Report: The AI Profit Machine, Wealth Concentration, and the Tension Between Social Production and Concentrated Ownership
Technical progress and the social question
AI can improve medical care and education and reduce repetitive work. Yet corporate revenue growth does not automatically deliver those benefits to everyone. The central questions are how knowledge, data and infrastructure are produced; who decides how they are used; and how benefits and costs are shared. These are central to A New Social Order in the Age of Consciousness. This report examines recent financial results and social risks in that light, without treating a company’s sales as proof of either general prosperity or inevitable impoverishment.
1. Investment forecasts and corporate results
Goldman Sachs Research estimates that global AI-related investment could reach about $1 trillion in 2026. This is a forecast for capital expenditure, not spending already completed or companies’ net profit. The researchers describe assumptions and some risk of double-counting.
In its official fiscal third-quarter report, Broadcom recorded $16.7 billion in AI semiconductor revenue for the quarter ended August 2, 2026, up 221% from a year earlier and 54% from the prior quarter. Its $21.7 billion figure for the next quarter is company guidance. Revenue is not the same as profit, and neither figure measures the personal income of CEO Hock Tan or any particular shareholder.
NVIDIA reported $89.0 billion in Data Center revenue in the second quarter of fiscal 2027, up 117% year over year. This growth illustrates how chips and compute capacity have become strategic bottlenecks. It does not, by itself, establish the company’s net profit or Jensen Huang’s personal gain.
Meta’s second-quarter 2026 filing shows total revenue of $60.8 billion, up 28% year over year. Advertising revenue was $59.36 billion, up 27%; net income declined from the comparable quarter. Sales growth, profit growth and returns on AI investment must therefore be kept distinct, even within one company.
2. Socially produced knowledge and concentrated control
AI systems draw on university research, public and private investment, engineers, supply-chain labor, cultural works and some data generated across society. Not every user interaction becomes training data for every model, and companies make real contributions to building the technology. Still, the book’s central question remains: when society helps produce knowledge and bears some of the costs, why can authority over models, chips, compute, platforms and rules of use become concentrated in a few institutions?
Corporate stock is held by different people and institutions; subscriptions do not simply flow into CEOs’ bank accounts. A stronger critique examines control rights, contracts, workers’ and users’ contributions, and who decides how productivity gains are shared. An openly available model alone does not democratize power if the compute and infrastructure required to use it remain scarce and concentrated.
3. Work, livelihoods and public costs
The International Labour Organization estimates that one in four workers worldwide holds an occupation with some exposure to generative AI; clerical occupations have the highest exposure. Exposure is not the same as a confirmed layoff. AI may reorganize tasks and improve productivity, while particular deployments can also weaken security and bargaining power. Any claim about a specific week’s layoffs or wage cuts needs a documented case, date and causal link.
The International Energy Agency projects that global data-center electricity consumption could reach roughly 945 TWh by 2030 and notes that large loads are geographically concentrated. In some places, grid investment raises difficult questions about who pays; the IEA also examines potential affordability pressures. Higher household water or electricity bills and local housing prices are not automatic outcomes of every data center. They depend on contracts, regulation, local capacity and public decisions.
4. A direction for a new social order
The answer is neither to reject the technology nor to replace corporate concentration with an unaccountable state monopoly. In the framework of A New Social Order in the Age of Consciousness, change should be conscious, gradual and open to revision: transparent energy and data contracts; meaningful participation by workers and local communities; privacy and creators’ rights; public oversight of foundational infrastructure; diverse social, cooperative and public ownership arrangements; and a fair social share in productivity gains. Shorter working hours, livelihood security and broad access to knowledge should become measures of progress.
The tension between the increasingly social production of knowledge and concentrated ownership does not dictate a single technological fate. It calls for democratic choices and institutions through which participation in producing knowledge becomes participation in power.
Sources
- Goldman Sachs Research, global AI investment forecast, 2026
- Broadcom, fiscal Q3 2026 results
- NVIDIA, fiscal Q2 2027 results
- Meta, Q2 2026 results
- ILO, Global Index of Occupational Exposure to Generative AI, 2025
- IEA, Energy and AI
- Hamid Akhavi, A New Social Order in the Age of Consciousness, Volume One; conceptual framework for this analysis.
