CATL

CATL's Yibin facility — the world's largest battery production site — faced mounting pressure as rapid capacity expansion drove up energy consumption and carbon emissions across all three scopes. The electronics and energy storage sector faces intense scrutiny over lifecycle emissions, and CATL's scale meant that inefficiencies compounded quickly: high Scope 1 and 2 emissions from direct operations, and a growing Scope 3 footprint tied to an extensive supplier network. Without systematic intervention, rising manufacturing costs and regulatory exposure around carbon disclosure would become structural liabilities as the facility scaled further.

CATL deployed an AI-driven energy transformation program centered on predictive ML models that continuously optimize energy consumption across the Yibin site's production processes. The system integrates with an on-site micro-grid combining photovoltaic generation and battery storage, allowing real-time AI-directed switching between renewable and grid power based on demand forecasting and production schedules. Process innovation ran in parallel, targeting direct emissions reductions at source. Critically, the program extended beyond the factory fence: CATL embedded low-carbon product R&D and a structured supplier carbon neutrality certification program to address Scope 3 emissions — engaging 13 upstream suppliers in measurable decarbonization commitments. This end-to-end approach was recognized by the World Economic Forum's Global Lighthouse Network in January 2026.

The Yibin initiative delivered a 56% reduction in overall carbon footprint — one of the most significant single-site industrial decarbonization outcomes reported. Specific scope-level results include:

• 16% reduction in annual Scope 1 & 2 emissions (direct operations and purchased energy)
• 45% reduction in Scope 3 emissions per unit of output (value chain)
• 13 suppliers achieved carbon neutrality certification through CATL's supplier program

The per-unit framing of the Scope 3 figure is significant: it means emissions intensity improved even as production volumes grew, demonstrating that decarbonization scaled with the business rather than being offset by expansion.

• AI energy optimization delivers the most impact when integrated with physical infrastructure — predictive ML alone has limited leverage without the micro-grid and renewable storage layer to act on its outputs.
• Scope 3 reductions require supplier engagement programs with formal certification pathways, not just procurement policy changes.
• Measuring Scope 3 on a per-unit basis rather than absolute terms provides a fairer benchmark for growing manufacturers and is more defensible under carbon disclosure frameworks.
• Lighthouse Network recognition can accelerate internal buy-in and external credibility for industrial AI programs at this scale.
• Battery and electronics manufacturers face outsized reputational and regulatory risk from carbon exposure — early investment in AI-driven energy systems provides compounding returns as disclosure requirements tighten.