Platforms for rapid customization To meet demand for air conditioners with climate-specific customizations, Haier in China deployed a product performance prediction model that simulates a design’s capacity based on customer specifications. The company then uses a machine learning algorithm to maximize it, reducing design cycle time by 49% and prototype verification cost by 77% per model.52 Simulating cost performance to reduce design risk and improve bid competition Midea in China did this is with a “one-click” product configuration tool that identifies the most cost- effective configuration out of millions of permutations. The site also developed a digital platform that auto- validates the manufacturability of designs with finite simulation algorithms that generate a comprehensive report, reducing design-for-manufacturing-assembly meetings from one day to 10 minutes. This tool relies on embedded design and feature recognition algorithms to relate new simulated designs to historically documented ones, all integrated into Midea’s in-house developed PLM platform that makes the full ecosystem of value chain data accessible, accelerating speed to market and cost competitiveness without sacrificing quality.53 Driving supply chain agility through intelligent resilience hubs 3.2 The COVID-19 pandemic inspired many to rethink their supply chains, realizing just how ill-prepared their organizations were to absorb costs from record- breaking disruption.54 However, Lighthouses tell a different story: 85% of them experienced revenue dips of less than 10% even at the height of the pandemic – while only 14% of other manufacturers could say the same.55 Since 2020, Lighthouses have implemented an average of five resilience- focused use cases, prioritizing supplier and customer connectivity and integrated planning.56More recently, Lighthouses are maturing solutions to manage the complexity of their fragmented and growing operating bases, especially when such value chains are exposed to market volatility. Solutions such as machine-learning “ensemble” models for demand forecasting and GenAI for commodity pricing and procurement stability are often orchestrated by resilience hubs that centralize risk management. They improve transparency across each node – from suppliers to end customers – and drive integrated decision- making for operational continuity and supply chain performance through disruption.EXAMPLE EXAMPLEHaier MideaSites face growing demands for rapid product customization and must deliver stringent requirements, complex configurations and smaller batch operations. To address this, companies such as Midea and Haier have invested in the central capabilities and digital tools that re-assign responsibility for product introduction tasks and decisions back to sites, improving agility and responsiveness to emerging opportunities. When data platforms are integrated across the end- to-end lifecycle of a product, valuable information can be fed back into early design stages to ensure optimal cost performance. Integrated advanced digital tools, such as automated product configuration and manufacturability validation platforms, utilize algorithms and historical data within a PLM system to streamline customization, reduce validation time and bid confidently on complex design requests. Since 2020, Lighthouses have implemented an average of five resilience- focused use cases, prioritizing supplier and customer connectivity and integrated planning. Global Lighthouse Network: The Mindset Shifts Driving Impact and Scale in Digital Transformation 27

Global Lighthouse Network 2025