Can the World Economy Grow Forever? How Global Growth Challenges Have Shifted Over the Past Decade

1. Historical Recap of World’s Economic Growth

The growth of World’s economy and population during the last 2000 years was not proportional and can be attributed mainly to the last 275 years. From the year 0 to 1750, Figure 1 shows world’s real GDP per capita grew only 60%; in the period from 1750 to 2010, it increased by over 900%. The growth in output was accompanied by the growth in population (Figure 2), which increased more than 900% during the last 275 years, compared to 400 % from 0 to 1750. As a result, the world’s real GDP increased by more than 800 times in the last two millenniums. This accelerated growth in recent 275 years was not a single event but a series of overlapping technological waves, or four industrial revolutions (IRs), each building on the last.

• IR1 from 1750 to 1830, when steam engine, cotton gin and early railroads were invented. It took more than 150 years for IR1 to have its full impact on US economy.
• IR2 from 1870 to 1900, when electric light, internal combustion engine, fresh running water to urban homes, telephone, phonograph etc. were invented. Full effects of IR2 on economy were seen in 100 years.
• IR3 from 1970s to 2010s, this revolution was built on the semiconductor, the internet, and mass computerization. As shown in Figure 3, the percentage of worldwide users grew from just over 15% in 2005 to 40% by 2014. Today, that number has soared to nearly 70%, transforming the internet from a niche tool into the foundational data and cloud infrastructure for the entire global economy.
• IR4 from 2020s to Present, this revolution is driven by Artificial Intelligence (AI) as a new General Purpose Technology (GPT). Unlike steam or electricity, AI is a GPT applied to cognition and knowledge itself. This wave runs on the data, compute, and connectivity infrastructure built by the Digital Revolution (IR3). It is already moving from a predicted "next big thing" to a measurable economic and scientific force. The timeline for discovery has collapsed. In drug development, AI platforms are now reducing the time to develop new therapeutic candidates from 5-6 years to as little as one year, with several fully AI-generated drugs entering Phase IIa clinical trials. In materials science, automated facilities like Berkeley's "A-Lab" now use AI to propose new stable compounds, which robots then immediately synthesize and test, creating a closed-loop discovery engine. In neuroscience, new AI-powered platforms are standardizing and analyzing data from tens of millions of brain cells, unlocking insights into diseases like Alzheimer's at an unprecedented scale. Economic data from mid-2025 suggests that the generative AI boom is a significant factor in the post-pandemic productivity surge, with aggregate labor productivity growth (2.16% annually) running significantly above its pre-pandemic trend (1.43%).

As a result of the four IRs mentioned above, the nature of challenges for world economic growth has changed significantly. While consumption of crucial resources and our impact on planetary systems (like the climate) are testing Earth’s capacity, the path for future growth has also transformed. Further growth no longer depends on consuming more, but on decoupling - using new technologies like AI and clean energy to drive resource efficiency and create new value while reducing our environmental footprint. However, as Figure 4 illustrates, the benefits of this growth have been profoundly unequal, with nations like China and Korea experiencing explosive late-20th-century growth while others have lagged - a divergence that the new AI-driven revolution will inevitably challenge and reshape. These challenges are discussed in the next section.

2. Challenges to Future Growth: From Scarcity to Transition

Past economic growth was driven by two factors:

• population growth,
• per capita GDP growth.

This model was fueled by a direct and compounding consumption of natural resources, which has pushed humanity's demand beyond Earth’s sustainable bio-capacity.
However, the nature of this challenge has fundamentally shifted. For further economic growth, use of energy, land and water causes the main challenges, which are discussed below.

Energy: The Challenge of Transition, Not Scarcity

The Old Challenge (c. 2015): The dominant view from a decade ago equated rising energy use with inevitable climate catastrophe. This perspective was rooted in historical trends, like those in Figure 5 and Figure 6, which clearly showed that rising consumption was overwhelmingly dependent on fossil fuels. The proposed solution was "steady state consumption" or radical efficiency.

The New Reality (2025): Today, that view is obsolete. The challenge is no longer energy use itself, but the carbon emissions from its generation. Driven by a near 90% drop in the cost of solar energy since 2010, the new paradigm is one of "electrify everything." This model promotes increasing clean electricity consumption to decarbonize transport (EVs), heating (heat pumps), and industry.

The New Bottleneck: The primary bottleneck is no longer a scarcity of energy but the speed of deployment. The new challenges are:

• Grid & Storage: Building the massive transmission lines and battery storage required to manage the intermittency of renewables.
• Critical Minerals: Securing the supply chains for lithium, copper, and cobalt needed for this new infrastructure. This is not just a mining challenge, but a new geopolitical bottleneck, creating complex new dependencies and potential for conflict.
• Investment: Mobilizing the capital to build out this new system.

Land: A Challenge of Yield, Not Just Area

The Old Challenge (c. 2015): The prevailing model, supported by data up to 2012 (as in Figure 6), showed global agricultural land use constantly expanding, suggesting we were nearing Earth's capacity. With productivity assumed to be relatively static, feeding a growing population seemed to require massive deforestation, locking growth and emissions together. The proposed solution was "steady state consumption."

The New Reality (2025): This view is now outdated. The core challenge has shifted from a hard limit on area to the efficiency and intensity of its use.

• "Peak Land" Has Been Reached: Recent FAO data (see Figure 7) indicates that total global agricultural land use has plateaued and is even slightly declining. We are successfully "decoupling" food production from land use - producing more food from less land.
• The Technology Revolution: The "static productivity" assumption was incorrect. The new model is one of technological intensification. This includes:
• Precision Agriculture: Using GPS, AI, drones, and IoT sensors to apply water and fertilizer with pinpoint accuracy, dramatically increasing yields and reducing waste.
• Genetic Engineering: CRISPR and advanced GMOs are creating crops with higher yields, better drought resistance, and less need for pesticides.
• New Farming Models: Vertical farming, while energy-intensive, can produce 10-20x the yield of a traditional farm using up to 95% less land and water, completely decoupling food production from arable land.
• Alternative Proteins: Plant-based and cultivated meats are vastly more land-efficient than traditional livestock, which uses a majority of the world's agricultural land.

The New Bottleneck: The constraint is no longer physical land. The new bottlenecks are social and financial: the speed of regulatory approval for new genetic technologies, the challenge of convincing millions of individual farmers to adopt capital-intensive precision agriculture, and the capital investment for new farming models.

Water: A Challenge of Energy and Management, Not Scarcity

The Old Challenge (c. 2015): The analysis, based on 2011 data, was one of scarcity. Conventional water supplies (rivers, aquifers) were at their limit. Unconventional sources like desalination were viewed as non-viable, costing "ten times more" than conventional water.

The New Reality (2025): This cost analysis is obsolete, as it missed two key technological shifts.

• The Energy-Water Nexus: The challenge of water is now an energy problem. The "10x" cost of desalination has collapsed. Driven by hyper-efficient reverse osmosis (RO) membranes and, most critically, the 90% drop in solar energy costs, large-scale desalination is now cost-competitive. Modern plants produce fresh water for as little as $0.40 - $0.50 per cubic meter. If you have cheap, abundant clean energy, you can have cheap, abundant fresh water.
• The Reuse Revolution: The old model overlooked advanced water recycling (potable reuse). Modern filtration and advanced oxidation processes can reliably and safely turn wastewater into high-quality drinking water. This creates a circular, local, and drought-proof water supply.
• Smart Management: In agriculture (the largest water user), precision irrigation and AI-driven water management are slashing waste.

The New Bottleneck: The problem is no longer a hard limit on supply. The new bottlenecks are:

• Capital Investment: Building the massive infrastructure for large-scale desalination and water recycling plants, which requires immense political will and stable public-private partnerships.
• Energy: Securing the vast amounts of (preferably clean) energy needed to power them.
• Public Perception: Overcoming the "toilet-to-tap" stigma for potable reuse, despite its proven safety.

3. Increase in Resource Productivity as a Source of Future Growth

The Stalling Engine 10 Years Ago (c. 2015): A decade ago, with hard physical limits on resources becoming clear, it was evident that future growth would have to come from productivity. This was viewed through two limited lenses: near-term incremental gains (like building insulation) and long-term technological development. The major concern, highlighted by analysis at the time (Jones, 2005), was that this long-term engine was stalling. The "burden of knowledge" had become too great; science was so vast and specialized that it took innovators decades to reach the frontier.

The AI-Driven Solution now (c. 2025): Today, the Fourth Industrial Revolution is providing a direct solution to that 2015 challenge. Artificial Intelligence is acting as a General Purpose Technology for cognition itself. It functions as a "co-scientist" that can synthesize the entire corpus of human knowledge, breaking down the research silos that defined the 20th century. This is already moving R&D from a slow, linear process to a rapid, exponential one, as seen in AI-driven drug and materials discovery. The "burden of knowledge" is being lifted.

The New Physical Bottleneck: This transformation has solved the primary cognitive bottleneck, only to reveal a new, critical physical one. The primary bottleneck on AI is no longer human knowledge, but energy and cooling.

• The On-Earth Bottleneck: Global data centers already consume over 2% of the world's electricity, and this demand is projected to surge. Forecasts for 2030 (see Figure 8) show a dramatic range of possibilities. This is not a failure of forecasting, but a "map" of our potential choices. The vast gap between the "Best case" (the green line) and the "expected" scenarios is the entire challenge. It represents the difference between aggressive investment in efficiency and new policy, versus a "business as usual" approach. The "expected" models (Andrae 2019, Belkhir/Elmeligi) project a rise to approximately 2,000 billion kWh/a, creating an unsustainable strain on local power grids and consuming vast amounts of water for cooling.
• The Next Frontier: The Cloud Above the Clouds: While terrestrial solutions like nuclear fusion or enhanced geothermal show long-term promise, they do not inherently solve the secondary bottleneck: the immense water and land use required for cooling. This terrestrial bottleneck is so severe that it is pushing the next logical step: moving gigawatt-scale data centers to space. This solution, highlighted by industry leaders at Amazon and OpenAI, is based on two immense advantages:
• Inexhaustible Energy: In orbit, solar panels receive 24/7, unobstructed sunlight, providing a "base load" power source.
• Perfect Cooling: The -270°C vacuum of space is a perfect "refrigerator," allowing heat to be passively radiated away without needing a single drop of water.
• The New Flywheel: A Space Economy: This vision faces significant challenges, primarily the non-negotiable laws of physics: high latency. This makes orbital data centers unsuitable for real-time applications like financial trading or streaming. These are not dealbreakers, but rather market-defining constraints. The first applications, already being deployed for "in-orbit edge computing," will be for "deep compute" - massive, non-urgent tasks like foundational AI model training, climate simulation, and scientific discovery. This provides the blueprint for the long-term "AI-Energy Flywheel." This model involves building space-based solar farms to power data centers in situ, moving a massive industrial load off-planet. This would not only create a self-sustaining space economy but would finally decouple high-end computation from Earth's energy and resource limits.

4. Conclusion

The invention of the printing press in 1430 changed how humanity shared knowledge, fueling the industrial revolutions. For centuries, this growth was built on a model of direct resource consumption, leading to the 2015-era conclusion that we were testing Earth's bio-capacity and that "steady state" consumption was the only path forward. This perspective assumed that our primary limits were the physical scarcity of Earth's resources and the cognitive scarcity of the human mind.

Today, the Fourth Industrial Revolution has made that entire framework obsolete. The challenges of growth have not been eliminated, but they have been transformed.

As this analysis has shown, the problem is no longer scarcity, but transition.

• For Energy, the limit is not a lack of power, but the political, financial, and geopolitical speed at which we can deploy abundant, cheap clean energy.
• For Land and Water, the limits are not a lack of physical area or supply, but the social and capital speed atwhich we can deploy technologies that decouple food and water from their traditional resource constraints.
• For Innovation, the primary bottleneck is no longer the "burden of knowledge." AI is solving the problem of human specialization and accelerating R&D.

This has solved the 20th century's cognitive bottleneck, only to reveal the 21st century's true physical bottleneck: the massive energy and cooling required to power this new global intelligence. This final challenge, however, is not a wall, but a doorway.

The solution is to move this new, heavy-industrial load off-planet. By building "the cloud above the clouds," we can power gigawatt-scale data centers with 24/7 orbital solar energy and cool them in the vacuum of space. This creates the "AI-Energy Flywheel" - the long-sought-after, profitable industry that can finally kickstart a self-sustaining space economy.

This optimistic framework must, however, contend with the Jevons Paradox - the economic principle that as efficiency increases, total consumption often rises. Why should this time be different? The answer lies in the nature of the transition. Past efficiencies (like the steam engine) made a scarce resource (coal) cheaper, encouraging wider use. Today’s transition is not just an efficiency gain; it is a system-level substitution (from finite fossil fuels to abundant renewables) and a cognitive multiplier (AI) that, for the first time, allows for growth in value (digital, scientific) that is structurally decoupled from material inputs.

Therefore, the answer to the question "Can the world economy grow forever?" has fundamentally changed. A decade ago, the answer was no; growth was limited by Earth's finite resources. Today, by decoupling our industries through clean energy and our cognition through AI, the answer is yes - because quantitative growth is no longer tethered to the limits of our planet.