0 cumulative citations
View corpus contextAI investment by itself does not deliver short‑run GDP gains across advanced economies, but paired with substantial renewable energy capacity it is associated with faster growth; AI adoption initially boosts energy use, yet shows an inverted‑U as efficiencies emerge beyond a threshold.
Can Artificial Intelligence Drive Sustainable Growth? Empirical Evidence on the AI–Energy–Growth Nexus in Advanced Economies
openalex
correlational
low evidence
7/10 relevance
Summary only
summary available; pdf_status=paywall
DOI
Source
PDF
Across G7 countries plus China and South Korea (2010–2025), AI investment alone is not associated with short-run GDP gains, but when combined with greater renewable energy capacity it correlates with higher growth; AI adoption raises energy demand initially, with an inverted-U pattern suggesting efficiency gains at higher adoption levels.
Citation observations
Cumulative provider counts captured on specific dates; providers are never combined.
This study examines the short-run relationship between artificial intelligence (AI), re-newable energy, and economic growth across the G7 countries, China, and South Korea over the 2010–2025 period. Motivated by the ongoing debate on whether AI-driven digital transformation can coexist with environmental sustainability, the analysis integrates technological and energy-economics frameworks. Using panel data and the Fixed Effects (FE) estimator with Driscoll–Kraay robust standard errors, four models (A1–A2–B1–B2) are estimated to explore how AI investment affects economic growth and energy demand. The results reveal that AI investment alone does not significantly enhance short-run economic growth; however, its interaction with renewable energy capacity yields positive and significant effects, confirming the moderating role of sustainable energy infrastructure. Conversely, AI development initially increases energy demand due to the expansion of data-driven infrastructure, but a non-linear (inverted U-shaped) relationship suggests that efficiency improvements emerge beyond a certain adoption threshold. Financial development and energy prices also play significant roles in shaping energy consumption dynamics. Overall, the findings indicate that AI-driven growth and energy efficiency are complementary in the presence of strong renewable capacity and innovation systems. The study provides empirical evidence for integrating AI policies with green energy strategies to foster sustainable digital transformation.
Summary
Main Finding
AI investment does not significantly boost short-run economic growth by itself across the G7, China, and South Korea (2010–2025). However, when combined with greater renewable energy capacity, AI investment has a positive and significant effect on growth. AI development raises energy demand at early adoption stages, but an inverted U–shaped relationship indicates that after a threshold of adoption/efficiency, AI contributes to lower energy intensity. Financial development and energy prices materially influence energy consumption dynamics.
Key Points
- Sample and scope: G7 countries + China + South Korea, annual panel 2010–2025.
- Estimation: panel Fixed Effects (FE) with Driscoll–Kraay robust standard errors to account for heteroskedasticity, serial correlation, and cross-sectional dependence.
- Model set-up: four models (A1–A2–B1–B2) examining AI investment effects on (a) short-run economic growth and (b) energy demand; models include interaction terms and non-linear (AI^2) specifications.
- Growth result: AI investment alone → not statistically significant for short-run growth; AI × renewable capacity → positive, significant moderator effect.
- Energy demand result: AI development → initial increase in energy consumption (expansion of data centers, digital infrastructure); AI^2 indicates inverted U–shaped relationship → beyond an adoption/efficiency threshold, further AI leads to lower energy demand per output.
- Other drivers: financial development and energy prices significantly affect energy consumption patterns.
- Interpretation: AI-driven digitalization can be environmentally compatible, but only when supported by robust renewable capacity and innovation systems; otherwise it tends to raise energy usage in the short run.
Data & Methods
- Data: country-level panel covering major advanced economies (G7) plus China and South Korea, 2010–2025. Key variables include measures of AI investment/development, renewable energy capacity, GDP/growth indicators, energy consumption, financial development, and energy prices. (Specific variable definitions and data sources not detailed in summary.)
- Econometric approach:
- Fixed Effects (FE) panel estimator to control for time-invariant country heterogeneity.
- Driscoll–Kraay robust standard errors to address cross-sectional dependence, serial correlation, and heteroskedasticity in inference.
- Four model specifications: two focused on growth (A1–A2) and two on energy demand (B1–B2), incorporating interaction terms (AI × renewables) and quadratic AI terms to capture non-linearities.
- Identification: short-run associations inferred from within-country variation over time; interaction and non-linear terms used to test moderating and threshold effects.
Implications for AI Economics
- Policy coordination: AI policy should be integrated with renewable energy investments. Public and private support for green energy infrastructure amplifies the growth benefits of AI while mitigating short-run energy demand increases.
- Investment sequencing: Prioritize scaling renewable capacity and energy-efficient data infrastructure alongside AI deployment to reach the threshold where AI yields net energy-efficiency gains.
- Innovation systems: Strengthen R&D, human capital, and institutions that enable diffusion of efficient AI technologies to accelerate movement past the inverted-U turning point.
- Energy pricing and finance: Use energy pricing and financial-sector policies to shape demand patterns—financial development affects energy consumption responses to AI, so financial policy can be a lever for sustainable digitalization.
- Research agenda: Examine heterogeneous threshold levels across countries, disaggregate AI subfields (e.g., model types, data-center vs. edge compute), and explore longer-run growth impacts and distributional effects of AI–renewables complementarities.
Assessment
Paper Typecorrelational
Evidence Strengthlow — The analysis is based on observational country-level panel regressions without plausibly exogenous variation in AI investment, leaving results vulnerable to reverse causality and omitted-variable bias; the sample is small (≈10 advanced economies) and aggregate, increasing the chance that unobserved confounders or measurement error drive estimated associations rather than causal effects.
Methods Rigormedium — Authors use standard and appropriate panel techniques (fixed effects, Driscoll–Kraay SEs), explore interactions and non-linearities, and control for key covariates, which is methodologically sound for descriptive inference; however, they do not exploit stronger identification (IVs, lag/exogeneity tests, event variation), and robustness to alternative specifications, measurement choices, and dynamic panel concerns is not described.
SampleAnnual country-level panel of G7 countries plus China and South Korea (10 economies) spanning 2010–2025; main variables include AI investment (country-level), renewable energy capacity, GDP growth (short-run economic growth), national energy demand/consumption, financial development indicators, and energy prices.
Themesproductivity adoption
IdentificationCountry-level panel fixed effects (within) regressions with Driscoll–Kraay robust standard errors to account for cross-sectional dependence and serial correlation; models include interaction terms between AI investment and renewable energy capacity and a squared AI term to test non-linearity; controls for financial development and energy prices. No instrumental variables, natural experiment, or exogenous shock exploited.
GeneralizabilitySmall-N sample of mostly advanced economies limits applicability to low- and middle-income countries, Country-aggregate analysis masks within-country heterogeneity across sectors, firms, and workers, Short-run focus (2010–2025) may not capture long-run structural effects of AI, Potential measurement error in AI investment and renewable capacity across countries, Policy, institutional, and technological heterogeneity across countries may limit transferability, Findings may be sensitive to model specification, functional-form assumptions, and omitted variables
Claims (7)
| Claim | Direction | Outcome | Confidence & Evidence | Details |
|---|---|---|---|---|
| AI investment alone does not significantly enhance short-run economic growth. Fiscal And Macroeconomic | null_result | short-run economic growth |
Reading fidelity
high
Study strength
medium
|
n=144
|
| The interaction of AI investment with renewable energy capacity yields positive and significant effects on economic growth, confirming the moderating role of sustainable energy infrastructure. Fiscal And Macroeconomic | positive | economic growth (effect of AI × renewable capacity interaction) |
Reading fidelity
high
Study strength
medium
|
n=144
|
| AI development initially increases energy demand due to the expansion of data-driven infrastructure. Organizational Efficiency | positive | energy demand / energy consumption |
Reading fidelity
high
Study strength
medium
|
n=144
|
| There is a non-linear (inverted U-shaped) relationship between AI development and energy demand, suggesting efficiency improvements emerge beyond a certain adoption threshold. Organizational Efficiency | mixed | energy demand / energy consumption (non-linear effect of AI) |
Reading fidelity
high
Study strength
medium
|
n=144
|
| Financial development and energy prices play significant roles in shaping energy consumption dynamics. Organizational Efficiency | mixed | energy consumption dynamics |
Reading fidelity
high
Study strength
medium
|
n=144
|
| AI-driven growth and energy efficiency are complementary in the presence of strong renewable capacity and innovation systems. Fiscal And Macroeconomic | positive | complementarity between AI-driven growth and energy efficiency |
Reading fidelity
high
Study strength
medium
|
n=144
|
| The study provides empirical evidence supporting the integration of AI policies with green energy strategies to foster sustainable digital transformation. Governance And Regulation | positive | policy recommendation for integrating AI and green energy strategies |
Reading fidelity
high
Study strength
medium
|
n=144
|