Exploding Expectations: Thoughts on AI-Induced Explosive Growth

The Economist has commendably undertaken a complex thought experiment: What happens if Artificial Intelligence fundamentally changes the world economy? The authors have done a great job, assembling an impressive phalanx of economists and theories to rationalize the most spectacular expectations—that is, to substantiate them with rational arguments. These expectations include exploding GDP growth, a future where all income flows to capital owners, soaring interest rates, and an unstoppable demand for gigantic data centers.

Although the authors repeatedly hold their argument up to the light—scrutinising data and methods rather than headline claims—they remain captive to prevailing economic paradigms. Insightful as their analysis is, it still reproduces the structural weaknesses embedded in contemporary economic theory. Truly anticipating AI’s systemic economic impact requires surmounting three methodological hurdles that mainstream economics has yet to clear.

First: To understand the economy-wide effects of AI, we would need an economy-wide model. Yet, to this day, economics lacks a truly holistic model that can depict the economy as an interconnected system with all its feedback loops. Analyses often remain trapped in isolated silos, and The Economist's arguments are no exception.

Second: To determine the effects of an information technology like AI, we must be able to model the use of information within economic processes. But we cannot. Economics usually treats information only implicitly—hidden in prices, quantities, or decisions—instead of modeling it as an independent, explicit variable that can be operationalized.

Third: To understand the effects of AI, we need a suitable model for its use. Instead, AI is often mistakenly treated as just another factor of production, like a machine or a software program, rather than what it is at its core: a supplier of information that is then used in an economic process. This weakness is intrinsically linked to the second.

From this, a compelling conclusion follows: A well-founded and substantiated analysis of the economy-wide effects of AI usage requires a framework that is holistic, information-sensitive, and operationalizes AI as a supplier of information.

When we apply such a systemic approach to the The Economist’s thought experiment and its conclusions, we arrive at fundamentally different and far less speculative conclusions.

In the following, we will examine four of the main propositions from The Economist's article. Our approach will be systematic. For each proposition, we will first model the causal chain implied by the article's logic. We will then contrast this with the causal chain derived from a systems-based perspective, the so called nanoeconomic view.

Kenneth J. Arrow coined the term “nanoeconomic” in 1987, describing it as an “extreme version of micro” analysis that sharpens the lens to the level of individual transactions. We use the term to mean microeconomics at an atomic level—the study of a single economic actor’s decisions, actions, and transactions. This deep level of analysis allows us to connect economic variables, operationalize information explicitly within processes, and identify the true leverage points of AI technology. This approach brings us closer to understanding how an initial impulse—wherever it arises in the economy—cascades into systemic effects.

By applying this systemic logic to the propositions of the The Economist's article, we will arrive at our conclusions.

These are the four key expectations we will address:

1. The use of AI will enable annual economic growth of 20-30%.
2. The demand for gigantic data centers will continue to grow unstoppably.
3. Extraordinarily high real interest rates of up to 30% will emerge due to massive investment needs.
4. All income will flow to a small class of capital owners as labor income is displaced.

Expectation 1: 20–30% Annual Economic Growth

The causal chain of the Economist article can be summarized as follows:

AI achieves superhuman capabilities (AGI) → AI automates not only production, but also research & development → Returns from automation are reinvested into more AI research → AI improves itself in a positive feedback loop → Economic growth (GDP) explodes to 20-30%+ per year.

The vision of a 20-30% annual growth rate is intoxicating, but its justification is built on high-level, aggregate variables, making its internal logic difficult to verify. Moreover, The Economist's logic focuses solely on our ability to produce more, faster.

A different causal chain emerges, however, when we view the economy through a systemic nanoeconomic lens—one that connects producers and consumers, treats AI as an information supplier, and explicitly models the information dependency of every economic decision and process.

Within this framework, we can translate the abstract concept of "increased AI usage" into a tangible mechanism: a marginal improvement in the information supply (a ΔInfo) at the precise points where this very information is needed. This ΔInfo becomes the fundamental trigger of change, and it is the key to the nanoeconomic analysis of AI impact.

Applying the nanoeconomic view to the issue of growth, the following causal chain emerges:

AI usage increases → Producers are able to generate more durable and better products → Consumer’s need for repairs, accessories, and replacement purchases decreases → The total number of purchases declines → Orders for factories and their output decrease → Measured economic activity (GDP) stagnates or declines.

Let us explore the logic of AI's impact within this causal chain using a practical example. The bicycle is my ideal case study: as a triathlete, I know it inside out. It neatly helps us understand the economic principles and mechanisms through which AI’s initial impacts ripple through markets and the wider economy, shaping their long‑term trajectory.

Today's bicycle is a source of constant follow-up needs. The chain needs oil, the tires wear out, the gears need maintenance, and the frame might rust. Each of these problems creates a transaction: we buy oil, new tires, or pay for a repair. All these purchases contribute to the GDP. Now, imagine a vastly improved bicycle, thanks to AI‑driven design: Its frame is made from a robust composite material that never rusts. The drivetrain is a frictionless, sealed unit that needs no maintenance for 100 years. The tires are made from a material that barely wears down.

This new bicycle is a vastly superior product. It dramatically increases our real wealth and quality of life ( My personal experience can prove that…). We have more time, less hassle, and more money in our pockets. But from the cold perspective of the GDP, a disaster has occurred. The entire economy that thrived on the imperfections of the old bicycle—the repair shops, the manufacturers of chain oil and spare tires—loses its business. The number of transactions plummets.

If we apply this logic to everything—from clothes that repel dirt to houses that need no maintenance to bridges built for millennia.—we see a clear trend. AI enables the creation of "perfect" products that eliminate entire cascades of follow-up needs. This leads to a "Great Production Paradox": we develop the capacity for hyper-efficient production, but the demand for that production paradoxically shrinks. The economy of "less, but better" leads to a stagnation or even a decline in measured GDP, while our actual quality of life soars.

Short answer to the Economist’s expectation: The growth ideology shifts from metrics to real long term quality.

Expectation 2: Endless Demand for Giant Data Centers

The causal chain of the Economist article for this expectation can be summarized as follows:

AI capabilities increase exponentially → The need for computing power for training & operation explodes → A growing economy requires more digital & physical infrastructure → Massive investments in energy & data centers become necessary → The demand for data centers increases relentlessly.

The image of endlessly sprawling data centres is striking, but it confuses a temporary symptom with a lasting trend. Today’s hunger for data processing is a sign of our system's inefficiencies, not its future. For now, these centers are busy merely rectifying past mistakes.

A systemic nanoeconomic view reveals this dramatically alternative causal chain:

AI usage increases → Factories, products, and systems become more efficient → Fewer errors and problems → Less constant monitoring and data analysis is needed → Less need for data processing in external data centers → The need for external data centers decreases.

Let's return to our bicycle case to explore this chain and pay a visit to the bicycle factory. In its current, "unintelligent" state, we need to collect massive amounts of data to make it work. We have to monitor every machine for potential failures. We analyze supply chain data to predict delays. We track market data to forecast demand. We need the power of huge, centralized data centers to process all this information, precisely because the system itself is still dumb and prone to errors.

Now, consider the "smart" bicycle factory. The machines are self-calibrating and signal their maintenance needs weeks in advance. They especially report their issues to their producers to overcome these in later versions. The production process is getting so efficient and robust that defects are virtually eliminated. The need for constant, centralized monitoring and brute-force data analysis diminishes and might even collapse one day. Furthermore, as we established, the demand for the follow-up products—the chain oil, the spare tires—disappears. This means the entire data processing requirement for their respective supply chains, marketing, and sales also vanishes.

The more probable long-term trend is towards embedding intelligence where it is needed, just as printing machines moved from the central printing department to every desk. A smart bicycle doesn't need to constantly ask a data center what to do. Its intelligence is on board. The current boom in data centers is the fever of a system undergoing a transformation. A healthy, intelligent, and efficient economy will have a much lower temperature and, consequently, a much smaller need for these gigantic data-processors.

Short answer to the Economist’s expectation: Huge data centers are needed to cure our current inefficiencies, and together with AI, they will do this job so well that they are digging their own graves.

Expectation 3: Real Interest Rates Will Explode

The causal chain for this extraordinary expectation in the Economist article can be summarized as follows:

Explosive growth is anticipated → Future incomes seem guaranteed to be high → The incentive to save decreases while the propensity to consume increases → Simultaneously, the demand for investment capital increases massively → To incentivize saving, real interest rates must rise extremely.

The idea of exploding interest rates is based on the assumption that the AI-transformation requires unimaginably large capital investments. This overlooks a crucial point: the technology improves not only the end products, but also the means of production themselves. The systemic view based on a nanoeconomic framework helps to clarify this interdependence:

AI usage increases → Production becomes highly efficient → Machines as a product of this production also become better & cheaper → Costs for building and replacing factories decrease → The investment requirement per factory decreases → The overall economic demand for credit declines → Interest rates decrease or remain low, instead of exploding.

Let us look deeper using our bicycle factory. A factory itself is just a very complex product. The same AI-driven efficiency that gives us a maintenance-free bicycle also gives us a low-maintenance, hyper-efficient machine for making bicycles. When the machines that build the factory become smarter, cheaper, and more durable, the cost of building the factory itself plummets.

Furthermore, advances in material science, spurred by AI, could lead to the development of "multi-purpose-factories." Imagine a single, versatile production system that can, with a simple software change, produce a bicycle now, custom-made dishes after lunch, and clothing before the evening. The need to build separate, hyper-specialized, and astronomically expensive factories for every product line would disappear.

This leads to a dramatic increase in capital efficiency. We can achieve a far greater output with far less investment. If companies need to borrow significantly less money to build and maintain their production capacity, the massive demand for credit that would justify exploding interest rates simply never materializes.

This logic is so fundamental that we do not even need to speculate about the consumer's propensity to save in the AI age.

Short answer to the Economist’s expectation: Why borrow massively for factories that become exponentially cheaper and more efficient? The premise for exploding interest rates evaporates.

Expectation 4: All Income Will Flow to Capital Owners

The causal chain of the Economist article for this most dramatic expectation can be summarized as follows:

AI becomes capable of doing human labor more cheaply → Wages are capped by the cost of AI → Human labor is increasingly replaced by capital (AI) and becomes redundant → Almost all income flows exclusively to the owners of capital → People must become pure capital owners to have an income.

The fear of a future where a few "super-rich capitalists" own everything while the rest of humanity is rendered obsolete is perhaps the most frightening, but also the most linear—and therefore most improbable—of the *Economist's* predictions. It ignores the second-order consequences of the evolution of technology.

A systemic view based on a nanoeconomic framework reveals this alternative causal chain:

AI usage increases → Production becomes automated and smart → Smart machines take over human jobs making products ever more efficient → Smart machines themselves become ever more efficient and cheaper → Costs for building efficient factories decrease → The production systems become ever more accessible to all humans → People can afford small smart factories, alone or collectively (the transformation).

The process begins as described: the rational desire of consumers for cheaper products forces companies to automate, displacing human labor. This is the painful, short-term conflict. But the story ends differently.

As we've established, the smart bicycle factory itself becomes a product that is subject to the same laws of efficiency. It becomes radically cheaper and easier to operate.

Historically, a lot of capital was needed because production was complex and prohibitively expensive, involving significant costs for materials, machines, and energy. But what happens when the "system of production" is a versatile, affordable, 3D-printer-like machine that can be set up in a local community center? This is akin to the evolution of the printing machine. Eventually, they became compact, affordable, and ubiquitous – literally available on every desk in homes and small businesses.

The consequence is profound: The conflict between 'labor' and 'capital' doesn't end with a victory for one side; it simply becomes irrelevant. The entire production system transforms, shifting its logic from selling one's time for a wage to buy goods from a producer toward a 'smart local factory' that produces what's needed, cheaply and on demand.

Ownership of the means of production becomes obsolete. Instead, direct access to the means of satisfying one's needs is all that matters. Perhaps this "little factory" will be community-owned, effectively belonging to no single individual?

Short answer to the Economist’s expectation: Why would anyone take the risk and expend effort on building and owning a factory when anything can be produced cheaper and locally by others? The income shift becomes irrelevant.

Final Thought: The Systemic Implosion of Production Costs

The Economist correctly identifies that in a world of automated production, the only constraints are "sufficient energy and infrastructure." But the analysis stops there. Let’s look at this proposition with a systemic view.

The five fundamental costs of production—Material, Labor, Capital/Machines, Energy, and Information—are all on a trajectory of radical reduction.

• Information is already becoming abundant and cheap.
• Labor is being replaced by smart machines.
• Capital/Machines become vastly cheaper and more efficient, as they are themselves a product of these more efficient production processes.

This leaves Material and Energy. While we face challenges here, no law of physics prevents us from leveraging AI and our creativity to discover and generate the "ultimate, universal, ubiquitous material (termed u3)" – one that can be reconfigured to serve as metal, glass, rubber, textile, plastic, and virtually any other material.

The development of new, super-efficient energy sources is directly connected to these material innovations. What if these smart local factories were autarkic, utilizing local energy production systems that are durable and require minimal maintenance?

The final destination of this journey might be a world where the cost of producing almost anything approaches zero. Maybe this is a return to the way of life of our ancestral campfire, but instead of scarcity and toil, this time we are equipped with the tools of abundance.