EVE Energy

EVE Energy, a major lithium battery manufacturer, faced mounting pressure in a market defined by razor-thin margins and escalating quality expectations. At their Jingmen facility, maintaining consistent cell quality across high-volume production lines proved difficult — even minor process deviations in battery manufacturing can cascade into defects that compromise safety and performance. The competitive electronics and energy storage market demanded both higher quality consistency and lower per-unit costs simultaneously. Without real-time process visibility and automated quality diagnosis, defect detection lagged behind production, driving up scrap rates, rework costs, and conversion expenses across the facility.

EVE Energy's Jingmen plant undertook a broad digital transformation, deploying over 40 integrated solutions spanning AIoT, simulation modeling, large language models (LLMs), and deep learning-based computer vision. Sensors embedded across production lines feed real-time data into AI systems capable of detecting anomalies in cell formation, coating, and assembly processes as they occur. Deep learning models analyze visual and sensor data to identify defect signatures that would be invisible to manual inspection. LLMs were integrated to support process self-optimization — interpreting production data, surfacing root causes, and recommending corrective actions. Predictive maintenance models monitor equipment health continuously, reducing unplanned downtime. The deployment covered operations holistically rather than targeting isolated bottlenecks.

The Jingmen transformation delivered measurable improvements across quality, cost, and equipment performance:

• 52% reduction in defect rate, significantly improving cell yield and reducing scrap and rework across production lines
• 41% reduction in unit conversion costs, reflecting the combined effect of fewer defects, less rework, and optimized process parameters
• 88% average overall equipment effectiveness (OEE), indicating high utilization, strong performance rates, and minimal unplanned downtime

The initiative also earned Jingmen recognition from the World Economic Forum's Global Lighthouse Network in January 2026, validating the factory as a benchmark for Fourth Industrial Revolution adoption in battery manufacturing.

• Breadth matters: Deploying 40+ integrated solutions across operations — rather than isolated AI pilots — was key to achieving system-level gains in OEE and cost reduction.
• Combine AI modalities: Pairing deep learning for defect detection with LLMs for process reasoning creates a more adaptive quality system than either approach alone.
• AIoT as the foundation: Real-time sensor integration is a prerequisite — AI models are only as useful as the quality and latency of the data they receive.
• OEE as the integration metric: Tracking OEE alongside defect rate captures the full operational impact, not just quality outcomes in isolation.
• Battery manufacturing is a strong fit: The high-stakes, high-volume nature of cell production makes the ROI case for AI-driven quality control particularly compelling.