Ellen Brown
May 26, 2026

Part 1 of this “Abundance Paradigm” series discussed predictions that artificial intelligence and robotics will in the relatively near future produce an economy of extraordinary abundance – one in which most labor is automated. The contention of Elon Musk is that this development will require some form of government-issued “Universal High Income” (UHI) to provide the consumer demand necessary to keep the economy functioning in a world where machines do most of the work.

Based on those projections, I argued that if a UHI were to become necessary, it could not realistically be financed through taxes or debt alone, but would require some form of debt-free sovereign money issuance — a modern version of Lincoln’s Greenbacks. The usual objection to government-issued money is that it would drive up prices and devalue the currency due to “too much money chasing too few goods.” But in this case, we would have too many goods and not enough money to provide the consumer demand to move them off the shelves. A source of abundant new money would actually be needed to keep trade flowing.

Objections came thick and fast. Some critics saw the AI revolution not as liberation but as a technocratic nightmare: AI surveillance, programmable digital money and “smart cities,” centralized control systems, and a future in which most people will own nothing while a tiny elite owns the machines, the data, and even the government. Others challenged the underlying premises: Would AI really generate such extraordinary abundance? Would productivity rise enough to justify something like a UHI? Or is this simply another round of Silicon Valley hype detached from economic reality?

Those are legitimate questions that deserve serious consideration, serious enough to require more than one sequel to address them. But whether or not we approve of Elon Musk, Sam Altman, or the AI industry itself, the AI revolution is already underway, driven by forces far larger than any individual actor. Businesses want AI because it lowers costs and increases productivity. Governments want it because they view it as strategically essential. Consumers increasingly rely on it because it saves time and improves convenience. The genie is out of the bottle.

Commentators say the AI boom is unlikely to disappear even if parts of it are overhyped. Investment firms, technology analysts, and economists increasingly describe AI not as a passing fad but as a foundational technological transition comparable to the invention of electricity or to the internet itself. Even skeptical analysts who question short-term productivity claims generally acknowledge that businesses are rapidly reorganizing around AI-assisted production.

The question now is not whether AI should exist but how we can adapt to it without falling into economic collapse or digital feudalism.

AI is Challenging the Fundamentals of the Capitalist Model.

For centuries, industrial economies have depended on a productive cycle based on work for pay. People work for wages, wages create consumer demand, and demand sustains production. But if machines increasingly perform not only factory labor but office and laboratory work — drafting contracts, diagnosing disease, designing products, writing software, driving vehicles, conducting research — then labor income will steadily decline even as productivity rises.

That creates a paradox for the capitalist model: Who buys the products if fewer and fewer people earn wages from producing them?

Historically, technological revolutions created new forms of employment even as they destroyed old ones. The automobile displaced blacksmiths but created mechanics, highway engineers, gas stations, motels, and suburbs. Computers eliminated typists but generated software industries and millions of office jobs. But AI is not confined to one sector. It is predicted to take jobs across the board.

We are not at that stage yet. But China, the world’s largest manufacturing power, is getting close, and Chinese commentators are beginning to grapple with the issue.

China as Forerunner and Test Case.

In a July 2025 opinion piece in the South China Morning Post titled “As AI Replaces Workers, China Could Consider Universal Basic Income,” Tech Editor Zhou Xin writes:

“In the past, Chinese officials have rejected proposals to distribute cash to households, even when many families were clearly in need of support. But while the term universal basic income has yet to appear in any official Chinese policy documents, it may become less foreign in the coming years because of the increasing replacement of entry-level jobs by machines.

Advances in technologies such as artificial intelligence (AI) and automation are expected to render many traditional labour roles obsolete ….

While new technologies will create new job opportunities, these roles are often unsuitable for workers displaced from traditional sectors. The pace at which old jobs are eliminated also outstrips the creation of new ones, which could lead to significant structural unemployment.”

A March 2026 article in ThinkChina raised a related issue. In “When AI Replaces Workers, Who Pays the Taxes?”, Chinese entrepreneur Simon Lin asks if AI systems and robots perform an increasing share of productive work, where will governments obtain tax revenue? Lin’s proposal is to tax the companies that profit from automation. That would help finance the government, but it doesn’t solve the distribution problem. Consumers still need purchasing power. Henry Ford understood this a century ago, when he said he needed to pay his workers enough to buy the cars they produced.

Another article in ThinkChina, titled “Socialism and Universal Basic Income: Creating Happy Societies in the Age of the Knowledge Economy,” addressed this issue in 2020. The article summary states:

“… [T]he knowledge economy offers great potential for bettering the lives of people. But capitalism may not be the best route to take. Power in the hands of a few, income gaps, job losses and wage cuts in the digital age bear this out. Can China offer a third way as it seeks to marry socialism with a market economy? The West is already considering some proposals with a socialist bent such as the Universal Basic Income (UBI). Surely, proponents of socialism can think of even more revolutionary ideas.”

The article continues:

“… China has a substantial low-income demographic. 600 million people live on about 1,000 RMB per month, which is insufficient even for housing rent alone. What we have here is inadequate demand from those with spending power, coupled with a tremendous surplus of production capacity.…”

The author observes that knowledge, once created, can be reused repeatedly at close-to-zero marginal cost, and that the AI-driven “knowledge economy” grows exponentially. That makes it possible for social productivity to grow exponentially as well, eliminating want and greatly enriching material and spiritual life. But capitalism poses some serious constraints on that promising future:

“… [A]s the knowledge economy becomes increasingly “smarter” (AI-driven), the share of wage income in the total distribution of income will continue to decline, while investment returns will be a constantly growing piece of the pie. This means the lion’s share of society’s wealth will be swallowed by capital. In the long run, only jobs with wages lower than the cost of automation have any chance of being kept.… This means that wage levels are bound to be kept low, even to the point of being inadequate for feeding oneself and one’s family.”

he article concludes: “China should kick-start preliminary research on universal basic income (UBI), as soon as possible. … What is UBI, after all, if not an attempt to rise above capitalism?”

Resource Constraints: Energy.

China may need to consider some sort of UBI, but in the United States the biggest practical hurdles to AI abundance may not be political but physical. Where will the U.S. find sufficient resources to produce the goods?

Critics point to the enormous energy consumption of AI data centers, the water demands of cooling systems, the mining requirements for batteries and semiconductors, and the environmental costs of rapid electrification. Some large data centers consume millions of gallons of water daily for cooling. Communities near rapidly expanding facilities have already reported stress on local water systems, and public pushback is growing.

Elon Musk has argued that the water problem is basically an energy problem, noting that once you have enough energy, desalination becomes cheap and simple. His proposed energy solution is solar. At a July 2017 National Association of Governors meeting, he said, “If you wanted to power the entire U.S. with solar panels, it would take a fairly small corner of Nevada or Texas or Utah; you only need about 100 miles by 100 miles of solar panels to power the entire United States. The batteries you need to store the energy, to make sure you have 24/7 power, is 1 mile by 1 mile. One square-mile. That’s it.” Not that all this equipment would need to be in one place, but that shows the projected scale.

The chief constraints to rapid and broad-scale solar development are political and regulatory. The solution being pursued now is solar collection in space, where the sun never sets, massive amounts of energy are available, cooling the equipment is not a problem, and there are no regulatory constraints.

Solar is not, however, the only possible energy solution. Advanced fission and fusion technologies are also in rapid development, largely due to AI-assisted engineering.

Small modular nuclear reactors (SMRs), once largely theoretical, are now moving into commercial development. SMRs are factory-built, standardized systems small enough in some cases to be transported by truck and assembled on site. Supporters argue that modular manufacturing could dramatically reduce both cost and construction time compared to conventional nuclear facilities.

Fusion energy, long mocked as perpetually “thirty years away,” is also advancing. Experimental reactors are already generating plasma temperatures hotter than the core of the sun, while major advances in magnetics are steadily improving stability. The main challenge is that superheated plasma behaves chaotically inside reactors, but AI systems are being used to predict these disruptions and make adjustments before they occur.

That doesn’t mean limitless energy is just around the corner. But the assumption that civilization is approaching an unavoidable energy ceiling may be outdated. In fact AI itself is becoming a key tool in creating the next generation of energy systems needed to support AI-driven productivity.

Physical Resources for Batteries, Electrical Grids and Agriculture.

AI is actually becoming a primary tool for solving resource problems in general. Modern AI-driven systems are dramatically improving electrical grid efficiency, agricultural productivity, recycling systems, and battery management. Precision agriculture reduces fertilizer and water use while significantly increasing yields. AI-managed electrical grids reduce wasted energy. Robotics improve mining precision and materials recovery. Advanced recycling systems increasingly recover rare earth minerals and lithium-ion battery materials that were once discarded as waste.

Thus while AI uses more power, the efficiency it creates in the rest of the physical economy may actually lead to a net reduction in total global resource consumption.

Solving the Water Crisis.

Singapore’s NEWater program is the gold standard for wastewater recycling, turning sewage into ultra-clean, drinkable water. It has now successfully “closed the water loop,” making the island nation resilient against external water shocks.

AI data centers are also now pivoting away from evaporative cooling to water recycling. Modern “closed-loop chilling systems” allow data centers to operate with near-zero direct water consumption once the system is filled. New major projects are marketing themselves as “water-neutral” by using closed-loop cooling technology that recirculates water rather than evaporating it in cooling towers.

Some analysts argue that the location of data centers is wrong. Unused areas are available that have abundant water supplies, existing industrial zoning, and underutilized energy infrastructure. But for communities already under stress from data centers that probably aren’t going anywhere, my own proposal would be to drill for primary (juvenile) water for residential needs. Continuously generated deep in the earth and rising through faults, primary water offers a clean, renewable, locally tappable water source independent of the surface cycle, easily accessible with robotic drilling and abundant energy. The model has been proven primarily in Africa. See my earlier article here.

Wind Power.

Meanwhile, China has successfully launched the world’s first commercial underwater data center powered directly by offshore wind. The Shanghai project was completed for less than half the cost of an equivalent 24-megawatt land-based facility, and by using seawater for cooling, it is about 30% more efficient and cuts electricity consumption by over 22%.

However, underwater data centers were not pioneered by the Chinese. Microsoft’s Project Natick, a 2018–2020 trial off the coast of Scotland, was a technically successful test that demonstrated higher reliability and lower failure rates than land servers. But Microsoft announced it was abandoning the project in mid-2024.

In the U.S., a private company like Microsoft must negotiate with local utilities and typically must pay for its own grid upgrades, which can add years and millions of dollars to a project. In China, state-owned power companies provide special energy pricing and dedicated high-voltage lines for data center clusters. There are also regulatory hurdles in the U.S. and Europe, where complying with environmental regulations is a slow and costly process.

In China, by contrast, the government designates specific areas where environmental reviews and construction permits are fast-tracked specifically for “Green AI” projects. As a result, construction is often 30% to 50% faster than for their Western counterparts. The Chinese underwater data centers are part of a massive state-led industrial policy called the “East-to-West Computing Resource Transfer,” a highly coordinated top-down strategy that treats data centers as a critical national utility integrated directly into the national energy grid. Besides providing direct subsidies and grants, the Chinese government has built offshore wind farms specifically designed to plug into data center units. Placing the AI servers directly at the base of the wind turbines eliminates the energy loss and cost of transmitting power back to the shore.

This is another real-world example demonstrating the need for public investment in infrastructure, ideally through a national infrastructure bank, to fund projects that private markets find too risky or too expensive to build alone.

The Road to Creative Freedom or to Digital Feudalism?

The potential for AI/robotic productivity is promising, but it will not automatically benefit the public. Productivity has already risen dramatically over the past century, while wealth has concentrated at the top.

The future emerging around AI contains two radically different possibilities. One is a highly centralized technocratic system in which wealth and power become highly concentrated, while citizens are managed through digital currencies, surveillance, and algorithmic governance. The other is a civilization in which automation gradually liberates human beings from monotonous labor, shortens work time, expands access to education and creativity, and allows technological abundance to serve broad human flourishing rather than narrow financial interests.

Both futures are technologically possible. Which one emerges will be determined not by the machines themselves but by the political and monetary systems governing them.

Part 3 will examine what is probably the most emotionally charged issue involved in the AI revolution: digital money, central bank digital currencies, surveillance fears, and whether an AI-driven economy inevitably leads to a programmable financial control grid.

Originally published on ScheerPost.com.

Shared via Creative Commons.