AI industry clusters first undermine but eventually enhance China's ability to jointly cut agricultural pollution and carbon emissions, producing a clear U-shaped effect; technological progress mediates the turning point and impacts are stronger where marketization and industrialization are higher.

Main Finding

Artificial intelligence industry agglomeration (AIIA) has a U‑shaped relationship with agricultural pollution–carbon reduction synergy (APCRS) in China (2016–2024): at low-to-moderate levels AIIA is associated with worsening APCRS (higher marginal cost / weaker coordinated abatement), but beyond a turning point further agglomeration improves APCRS. Technological progress partially mediates this nonlinear effect; several institutional and structural factors alter the shape and timing of the U‑curve.

Key Points

• New metric: APCRS is constructed on a marginal-cost basis to capture the joint (coordinated) abatement of agricultural pollution and carbon emissions rather than single‑pollutant effects.
• Main pattern: AIIA → APCRS is U‑shaped. Early-stage agglomeration tends to reduce synergy (increase marginal abatement cost), while sufficiently advanced agglomeration enhances synergy.
• Mediation: Technological progress is a partial mediator—AI agglomeration improves technology, which helps reduce the marginal cost and shift outcomes toward the favorable part of the U.
• Moderators:
  • Agricultural socialized services and rural industrial integration mitigate the initial negative association (they buffer early adverse effects).
  • Higher agricultural labor productivity amplifies the curvature (strengthens the nonlinear pattern).
• Heterogeneity:
  • The positive rebound at higher AIIA levels is stronger in provinces with higher marketization and higher industrialization.
  • The U‑shaped relationship holds within grain strategic zones (i.e., the nonlinear pattern is robust across these agricultural policy contexts).
• Threshold dynamics: The beneficial phase (movement toward improved APCRS) is more likely to occur once public innovation investment and rural household income exceed certain thresholds.

Data & Methods

• Data: Chinese provincial panel (30 provinces) for 2016–2024.
• Key variable: APCRS — an indicator of agricultural pollution–carbon reduction synergy built from marginal-cost considerations (captures coordinated abatement efficiency).
• Explanatory variable: degree of AI industry agglomeration (AIIA).
• Empirical approach (summary of techniques used):
  • Panel regression frameworks to estimate the AIIA–APCRS relationship allowing for nonlinearities (identifying U‑shape).
  • Mediation analysis to test the role of technological progress.
  • Interaction/moderation analyses with agricultural socialized services, rural industrial integration, and labor productivity.
  • Heterogeneity tests across levels of marketization/industrialization and across grain strategic zones.
  • Threshold analysis to detect critical levels of public innovation investment and rural household income that change the dynamics.
• Robustness: results reported as robust across regional subsamples and conditional on external institutional/structural factors (as noted above).

Implications for AI Economics

• Nonlinear environmental impacts of AI agglomeration: Policymakers and analysts should not assume monotonically positive or negative environmental effects from AI clusters—benefits can require reaching scale, complementary institutions, and investments to materialize.
• Importance of complementarities: Public innovation spending, rural incomes, agricultural socialized services, and rural industrial integration are critical complements that hasten the transition from the adverse to the beneficial phase of AI agglomeration.
• Targeted industrial policy: Regions at early agglomeration stages need supporting policies (service networks, integration, income support, public R&D) to avoid a trap where nascent AIIA raises marginal abatement costs.
• Evaluation metric: Using marginal‑cost–based synergy measures (like APCRS) gives a more policy‑relevant view of joint pollution and carbon outcomes than single‑pollutant indicators; AI economics studies of environmental impacts should adopt such joint-cost perspectives.
• Spatial and institutional tailoring: The effectiveness and timing of AI cluster benefits for green agriculture depend on regional marketization and industrialization—policy design should be region‑differentiated.