US AI investment rises with economic growth, trade openness, renewable energy use and global integration, but energy-policy volatility shows a Goldilocks effect — some uncertainty spurs innovation, while high unpredictability undermines long-term AI investment.
Do energy policy uncertainty, trade openness, and renewable energy drive artificial intelligence investment? Evidence from the United States
Using wavelet-based quantile methods on US quarterly data (2013–2024), the paper finds that renewable energy use, GDP growth, trade openness and globalisation are positively associated with AI investment, while energy policy uncertainty has a nonlinear (Goldilocks) effect—moderate uncertainty encourages investment but high volatility deters it.
The growing incorporation of artificial intelligence into economic frameworks has prompted investigations on the macroeconomic and policy determinants influencing AI investment, especially in the United States. Despite its increasing significance, there is a paucity of empirical study examining the collective impact of trade openness, renewable energy use, economic growth, globalisation, and energy policy uncertainty on AI investment. This study utilises sophisticated nonlinear methodologies Wavelet Quantile Regression (WQR) and Wavelet Quantile Correlation (WQC) to analyse distributional asymmetries and time-frequency dynamics from 2013Q1 to 2024Q4. The findings indicate that renewable energy consumption, economic growth, trade openness, and globalisation positively influence AI investment, but energy policy uncertainty has a nonlinear effect, with moderate uncertainty fostering innovation but high volatility hindering long-term investment. These results highlight the significance of stable energy policy, continuous economic growth, and improved global integration for promoting AI development. The study emphasizes the need for integrated policies that connect energy transition, macroeconomic stability, and digital innovation to maintain the United States’ technical supremacy.
Summary
Main Finding
Using quarterly U.S. data (2013Q1–2024Q4) and wavelet-based quantile methods, the paper finds that renewable energy consumption, economic growth (GDP), trade openness, and globalization are robust positive drivers of AI investment. Energy policy uncertainty (EPU) has a nonlinear, distribution- and frequency-dependent effect: moderate uncertainty can spur short-term innovation/investment while high or sustained policy volatility deters long-run AI capital formation.
Key Points
- Dependent variable: AI investment (AI index). Explanatory variables: Renewable energy consumption (REN), GDP, Energy Policy Uncertainty (EPU), Globalization (KOF-style measure, GLO), Trade Openness (TO).
- Hypotheses tested:
- H1: GDP → + AI investment
- H2: REN → + AI investment
- H3: EPU → − AI investment (but nonlinearity expected)
- H4: GLO → + AI investment
- H5: TO → + AI investment
- Main empirical results:
- REN, GDP, TO, and GLO show positive associations with AI investment across many quantiles and across short-, medium-, and long-term frequency bands.
- EPU displays asymmetric/nonlinear effects across quantiles and scales: moderate or transitory increases in EPU can be associated with higher AI investment (interpreted as adaptive or opportunistic innovation), whereas high magnitude or persistent EPU reduces long-horizon investment.
- Methodological advantage emphasized: moving beyond symmetric, time-invariant regressions to capture both distributional heterogeneity (different behavior in low- vs high-investment regimes) and time-frequency dynamics (short vs long horizon effects).
Data & Methods
- Sample: United States, quarterly observations covering 2013Q1–2024Q4.
- Variables and sources (as reported in manuscript):
- AI: Investment in Artificial Intelligence (AI index; Maslej et al. 2025 / Our World in Data)
- REN: Renewable energy consumption (percent-equivalent)
- GDP: Real economic output (log-transformed)
- EPU: Energy Policy Uncertainty index
- GLO: Globalization index (multi-dimensional/KOF-style)
- TO: Trade openness (exports + imports / GDP)
- Data treatment:
- All series log-transformed to reduce heteroskedasticity.
- Lower-frequency data converted to higher-frequency using a quadratic match-sum-up approach when needed.
- Econometric approach:
- Wavelet Quantile Regression (WQR): combines quantile regression with wavelet decomposition to estimate relationships across the conditional distribution of AI investment and across time-frequency scales (short, medium, long run).
- Wavelet Quantile Correlation (WQC): assesses quantile-dependent correlations across scales.
- Rationale: WQR/WQC capture asymmetric responses, time-varying dependencies, and multi-scale effects that standard linear/specification-invariant methods miss.
Implications for AI Economics
- Policy design:
- Stable energy policy matters for securing long-horizon AI investment. Policymakers should minimize protracted uncertainty in energy regulation and carbon/price regimes to reduce option-value-driven investment delay.
- Renewable energy expansion supports AI investment by improving power sustainability, lowering operating costs for data centers/cloud infrastructure, and creating demand for AI-enabled grid management—linking decarbonization and digitalization policy agendas.
- Trade openness and international integration (globalization) facilitate technology diffusion and access to inputs (hardware, semiconductors, software), thus supporting AI capital formation—trade and digital-policy coordination is important.
- Industry strategy:
- Firms may opportunistically increase AI investment under transient policy shocks, but sustained policy volatility raises financing costs and discourages fixed-cost AI commitments (e.g., data centers, specialized R&D).
- Investments that tightly couple AI infrastructure to renewable energy (e.g., colocated data centers with renewables, AI for grid optimization) can capture complementarities highlighted in the study.
- Research directions:
- AI economics research should account for nonlinear and multi-horizon effects when studying investment drivers—wavelet-quantile tools are useful for this purpose.
- Further work could quantify threshold magnitudes/durations of EPU that flip its sign from supportive to detrimental, and explore sectoral heterogeneity (e.g., cloud providers vs. enterprise adopters).
- Strategic takeaway: Coordinated policy that combines predictable energy regulation, support for renewable capacity, and open trade links creates a more favorable macro-structural environment for sustained AI investment and for preserving technological leadership.
Assessment
Paper Typecorrelational
Evidence Strengthlow — The analysis is based on aggregate observational quarterly time series (2013Q1–2024Q4) and reports associations rather than causal estimates; potential endogeneity, reverse causality, omitted variables, and confounding from concurrent shocks (e.g., COVID, energy price spikes) are not addressed by exogenous identification strategies.
Methods Rigormedium — The authors apply sophisticated non-linear, time-frequency tools (WQR and WQC) which are well suited to exploring distributional asymmetries and frequency-specific dynamics, but methodological rigor is limited by the short macro time series, sensitivity to wavelet/quantile choices, and apparent lack of strategies to handle endogeneity, structural breaks, or alternative causal checks (IVs, natural experiments, placebo tests).
SampleQuarterly aggregate United States time-series from 2013Q1 to 2024Q4 (approximately 44–48 quarters), covering measures of AI investment (unspecified proxy), renewable energy consumption, real economic growth (GDP), trade openness, globalisation index, and an energy policy uncertainty metric.
Themesadoption innovation governance
IdentificationTime-series association analysis using Wavelet Quantile Regression (WQR) and Wavelet Quantile Correlation (WQC) to estimate distributional and time-frequency relationships between quarterly US AI investment and macro/energy variables (renewable energy consumption, GDP growth, trade openness, globalisation, energy policy uncertainty); no quasi-experimental or instrumental identification is reported.
GeneralizabilitySingle-country (United States) aggregate analysis — may not generalize to other countries or regions, National-level quarterly data — cannot capture firm-, sector-, or worker-level heterogeneity, Relatively short sample window (≈11–12 years) limits inference about long-run dynamics, Findings depend on proxies and index construction (AI investment, globalisation, EPU) which may not be comparable across studies, Results may be sensitive to major contemporaneous shocks (pandemic, supply-chain disruptions, energy crises) that are not fully disentangled
Claims (8)
| Claim | Direction | Confidence | Outcome | Details |
|---|---|---|---|---|
| Renewable energy consumption positively influences AI investment in the United States. Adoption Rate | positive | high | AI investment |
n=48
0.3
|
| Economic growth positively influences AI investment in the United States. Adoption Rate | positive | high | AI investment |
n=48
0.3
|
| Trade openness positively influences AI investment in the United States. Adoption Rate | positive | high | AI investment |
n=48
0.3
|
| Globalisation positively influences AI investment in the United States. Adoption Rate | positive | high | AI investment |
n=48
0.3
|
| Energy policy uncertainty has a nonlinear effect on AI investment: moderate uncertainty fosters innovation, whereas high volatility hinders long-term investment. Adoption Rate | mixed | high | AI investment |
n=48
0.3
|
| Wavelet Quantile Regression (WQR) and Wavelet Quantile Correlation (WQC) effectively capture distributional asymmetries and time–frequency dynamics in the relationships between macro/policy determinants and AI investment. Research Productivity | positive | high | distributional asymmetries and time-frequency dynamics of macro determinants' relationships with AI investment |
n=48
0.5
|
| Stable energy policy, continuous economic growth, and improved global integration are significant for promoting AI development in the United States. Governance And Regulation | positive | high | AI development / AI investment |
n=48
0.05
|
| Policymakers should pursue integrated policies linking energy transition, macroeconomic stability, and digital innovation to preserve the United States' technical supremacy in AI. Governance And Regulation | positive | high | preservation/promotion of US technical supremacy in AI |
n=48
0.05
|