Artificial intelligence (AI) is becoming one of the most important variables in global economic growth. From predictions by Silicon Valley optimists to calculations by mainstream economic institutions, whether AI can boost global economic growth from the current 2-3% to an "explosive" level of 20-30% has become a core topic of heated debate. This article combines historical economic growth trajectories, economic theoretical models, the latest investment and energy data for 2026, and current bottlenecks to conduct a systematic analysis.
Historical Perspective: From Stagnation to Accelerated Growth Paradigm Shift
1700 years ago, the global economy grew at an average annual rate of only about 0.1%, essentially stagnant.1700 years ago, the global economy grew at an average annual rate of only about 0.1%, essentially stagnant.
... Following the Industrial Revolution, technological breakthroughs such as the steam engine propelled the growth rate to 0.5% between 1700 and 1820, further reaching 1.9% by the end of the 19th century. In the 20th century, global output grew at an average annual rate of 2.8%. This long-term trend demonstrates that technological innovation, by increasing productivity and capital accumulation, has achieved a step-by-step leap in growth rates. AI is considered a general-purpose technology similar to or even surpassing the Industrial Revolution. Unlike previous technologies, AI possesses the potential for self-iteration, automating the vast majority of cognitive and physical tasks, thereby achieving an exponential acceleration of "labor accumulation." This contrasts sharply with historical "population accumulation": traditional growth relies on generational replacement, while AI "workers" can be rapidly replicated through investment. The Theoretical Mechanism of AI Explosive Growth Mainstream economic growth models, assuming that AI can effectively replace human labor, often predict explosive growth. Both semi-endogenous and exogenous growth models show that when the cost of AI is lower than human labor and the investment ratio is sufficient (e.g., more than 20% of GDP), the rapid accumulation of AI agents will form a positive feedback loop: automation increases output → reinvestment in more AI → productivity snowballs. Research by Epoch AI and others indicates that if an AI system can perform work equivalent to human labor at an annual cost of less than $15,000, and hardware efficiency continues to improve, the global economic growth rate may exceed 30%. In an optimistic scenario outlined in the World Bank's 2026 report, AI-driven productivity gains could propel global growth back to or even surpass the 2000s highs of the 2030s. The IMF also believes that AI investment has already significantly contributed to US GDP growth in 2026 and may contribute an additional 0.1-0.8 percentage points to global growth in the medium term. Key mechanisms include: task automation, improved single-task productivity, and the accelerated R&D itself (recursive self-improvement). Institutions such as Morgan Stanley predict that global growth will be around 3.2% in 2026, with AI capital expenditure being a major driving force. The Real Investment Wave and Energy Infrastructure Bottlenecks: By 2026, AI investment has moved from concept to large-scale implementation. Data center power consumption has become the most direct indicator. IEA data shows that global data center power consumption was approximately 485 TWh in 2025, and is projected to double to 950 TWh by 2030, accounting for about 3% of global electricity consumption. US data center power demand may increase from 80 GW to 150 GW between 2025 and 2028. McKinsey estimates that by 2030, AI-related data center infrastructure will require $5.2 trillion in investment, with technology hardware accounting for 60%. Giants such as Alphabet, Amazon, and Meta plan to invest over $350 billion in 2025-2026. On the energy front, renewable energy purchase agreements (PPAs) are surging, but grid bottlenecks, land constraints, and water consumption (some large data centers consume millions of gallons of water daily) pose real constraints. While these investments drive short-term growth, they could trigger capital shortages and rising interest rates. In a scenario of explosive growth, high returns will reduce the willingness to save, while infrastructure demand will push up borrowing costs, leading to higher long-term bond yields. This could, in turn, suppress asset prices, creating a complex dynamic equilibrium. **Job Transformation: Automation Risks and Cost-Disease Effects** The impact of AI on employment is not a simple replacement. Employment in automated tasks will be rapidly impacted, but opportunities will remain in non-automated fields (such as plumber jobs requiring physical dexterity or complex interpersonal interactions). Historical experience shows that sectors with rapidly increasing productivity tend to drive up overall wages through "Baumol Cost Disease," while wages in low-productivity sectors also rise, providing a buffer for displaced workers. Experts predict that even under a scenario of "rapid AI progress," the labor force participation rate may decline by 2050, but GDP growth will accelerate to approximately 3.5-4%. The Wharton model is more conservative: AI will increase productivity and GDP by 1.5% by 2035 and by 3.7% by 2075. China's advantages in robotics and embodied AI are particularly prominent. The combination of its manufacturing hardware strength and AI software planning is expected to give it a leading position in the integration of supply chains and the real economy. Capital Market Implications: Valuation vs. Macroeconomic Signals Silicon Valley's high valuations reflect bets on the long-term dominance of AI companies, but in the money market, explosive growth has not yet been fully priced in. Long-term bond yields are a key indicator: a significant rise suggests the market believes the overall economy will "explode"; high valuations for only AI companies are more likely a continuation of a normal growth cycle driven by specific technologies. Compared to the dot-com bubble, AI differs in its ability to accelerate the frontiers of knowledge. If AI can generate research ideas and break through scientific bottlenecks, its long-term improvement in living standards will far exceed that of the internet era. The Stanford 2026 AI Index shows that AI adoption is accelerating at a historic pace, and businesses and consumers have already reaped substantial value from it. Policy, Regulation, and Global Imbalances The growth potential of AI is unevenly distributed across different countries. Advanced economies, with their digital infrastructure and human capital, are better positioned to capitalize on this opportunity, while emerging markets need to bridge the digital divide. China has emphasized the deep integration of physical industries and AI, as well as the development of robotics, at events such as APEC, and is actively making strategic deployments. Globally, a balance needs to be struck between innovation incentives and regulation: data privacy, ethical standards, and restrictions on robot deployment could all pose bottlenecks. Fiscal policy should focus on retraining, infrastructure, and R&D subsidies. The combination of energy security and AI investment could become a new engine for growth. Outlook: A Future of Optimism and Caution AI has the potential to significantly boost global productivity and economic growth, but explosive growth hinges on multiple conditions, including self-improvement, cost reductions, and breakthroughs in bottlenecks. Under the baseline forecast, AI will contribute stable growth momentum from 2026 to 2030; an optimistic scenario could lead to historic high growth; a pessimistic scenario may simply be another technological wave, accompanied by significant disruption but limited overall output improvement. Policymakers, businesses, and investors should closely monitor actual productivity data, energy supply progress, labor market adjustments, and bond market signals. The AI era is not an inevitable utopia, but rather a window of opportunity that needs to be proactively shaped. Its inclusive potential can only be maximized through technology governance, investment in talent, and international cooperation.