美的科技背后的探路者 站在无人区里炼金

Summary: The article's core is Midea's long-cycle trial and error in uncharted hard-tech territory, achieving multiple 0-to-1 breakthroughs and forming a reusable R&D methodology.

1. After five years and dozens of iterations, the RV reducer achieved an 8,000-hour constant-load lifespan, 3,000 reciprocating impact validations, and complex curved-surface machining accuracy deviations controlled below 2-3 microns; it entered mass production and won a technology breakthrough award.

2. Surface acoustic wave technology for microwave ovens solves uneven heating; the project restarted in 2021 and achieved a breakthrough over three years. In 2024, when mold opening encountered consistency failures, 20 to 30 people worked overnight to validate nearly 40 schemes, ultimately winning the National Science and Technology Progress Award.

3. Blue oxygen for washing machines uses electrocatalytic oxidation to generate 2.8V hydroxyl radicals, enabling sterilization, stain removal, and color protection without chemical additives, and uses solid-state proton conduction technology to adapt to different water quality.

4. Methodologically, it emphasizes forward design, three generations, unrestricted trial and error, and platformization and modularization; refrigerator platforms were reduced from more than a dozen to 5, the development cycle fell from 18 months to 7 months, and costs dropped by more than 10%.

5. AI is already used for material simulation and TRIZ efficiency improvement, and agent-based AI generates hundreds of millions of yuan in annual efficiency gains.

Summary: The most valuable aspects for brands are product R&D, user insights, cost, and overseas brand strategy.

1. In product R&D, Midea uses underlying technology to solve long-standing pain points: surface acoustic wave technology addresses 60 years of uneven heating in microwave ovens; blue oxygen enables additive-free sterilization, stain removal, and color protection; and second-generation plasma technology lowers costs and extends to air conditioners, floor scrubbers, dishwashers, and other full categories.

2. In user behavior observation, the refrigerator team moved the cooling system to the vertical beam for small-apartment sideboards to free up internal space; the North America team used 1,000 questionnaires to distill pain points around ice splashing when dispensing ice water and cup-size confusion, tested them in employees' homes, and launched the industry-first Ice&Water Autofill series.

3. Platformization and modularization bring cost and efficiency gains: refrigerator platforms were reduced from more than a dozen to 5, the development cycle was shortened from 18 months to 7 months, and costs fell by more than 10%; first-generation plasma cost 35 yuan, and the second generation landed across categories after overcoming cost and reliability challenges in complex environments.

4. At the brand and market level, North America products became a benchmark for the overseas OBM priority strategy; the C-end technology pool is migrating to AIDC liquid cooling, SST insulating interface materials, high-efficiency power supply and distribution, and robotics, supporting ToB brand expansion.

Summary: Sellers should watch demand changes, growth markets, partnership opportunities, and risk alerts.

1. Demand and opportunities: pain points such as uneven microwave heating, detergent residue and disinfectant irritation, small-apartment sideboard space, and ice splashing in North America have been addressed through technology and can be converted into category upgrades and overseas OBM opportunities; AIDC computing infrastructure, heavy-duty robots, medical equipment, humanoid robots, and embodied intelligence are Midea's expansion directions.

2. Partnership approaches: the RV reducer combines technology breakthroughs by the Central Research Institute, manufacturing industrialization by Jiya Precision, and complete-machine application validation by KUKA, and the product has entered mass production; Blue Orange Lab pursues differentiated expansion for external complete-machine customers such as KUKA, and a dedicated team has been set up to advance industrialization of lab results.

3. Learnings: use forward design and the three-generation relay to run development, reserve, and research in parallel; use platformization and modularization to reduce cost and improve efficiency, with refrigerator platforms reduced from more than a dozen to 5, the cycle shortened from 18 months to 7 months, and costs down by more than 10%; AI agents generate hundreds of millions of yuan in annual efficiency gains.

4. Risk alerts: good technology can fail when water quality changes, requiring solid-state proton conduction technology to create a globally universal module; before mold opening there was a fatal consistency failure, requiring batch validation; humanoid robots should prioritize industrial use, follow with commercial use, and explore home use, without blindly pushing immature products to market, as hardware stability and generalization remain a long journey.

Summary: Factory readers can absorb practical lessons in four areas: production design, mass-production consistency, digitalization, and business opportunities.

1. Production and design requirements: the RV reducer must pass 8,000 hours of constant-load life and 3,000 reciprocating impacts, with complex curved-surface machining accuracy deviations locked below 2-3 microns; the team tackled precision machining, manufacturing, testing, measurement, test evaluation, raw materials, and heat treatment item by item, showing that high-end core components impose extremely high requirements on processes and validation systems.

2. Mass production and consistency: Midea emphasizes shifting from technology breakthroughs to mass production and benchmarking consistency and quality stability; before mold opening for micro-steam-bake, dozens of trial units still showed large deviations and a fatal consistency failure occurred, and 20 to 30 people worked overnight with experts to co-create nearly 40 schemes and validate them batch by batch, showing that batch validation and exception handling are essential before mass production.

3. Design and platformization: refrigerators move the cooling system to the vertical beam to achieve center cooling, platforms were reduced from more than a dozen to 5, the development cycle was shortened from 18 months to 7 months, and costs fell by more than 10%, reflecting the approach of building large platforms and strong modules.

4. Digitalization and e-commerce implications: the Jingzhou factory built the world's first agent factory, and the Wuxi high-end factory is exploring an end-to-end factory brain; the Yutu inspection robot ran continuously for more than 14 days without human intervention, automatically charged, and performed EHS checks, while production-line screw fastening and parts sorting achieved success rates above 99%; AI agents generate hundreds of millions of yuan in annual efficiency gains.

Summary: Service providers can grasp hard-tech and AI+ trends, customer pain points, and reusable solutions.

1. Industry trends: Midea is expanding in high-end heavy-duty robots, RV reducers, AIDC computing infrastructure, medical equipment, humanoid robots, and embodied intelligence, emphasizing 0-to-1 bottleneck breakthroughs and unrestricted industrial research; AI is viewed as new electricity and new infrastructure, with industrial scenarios first, commercial use following, and home exploration later.

2. New technologies: high-precision and high-reliability RV reducers, engineering of surface acoustic wave technology, blue oxygen electrocatalytic oxidation, solid-state proton conduction, plasma, data center liquid cooling, SST insulating interface materials, high-efficiency power supply and distribution, as well as AI large models, TRIZ combined with AI, AI agents, and factory brains.

3. Customer pain points: supply chain risk from reliance on imports for core components; 60 years of uneven microwave heating; detergent residue, disinfectant irritation, and high energy consumption of steam; different national water quality causing technology to fail; high plasma costs and insufficient reliability in complex environments; ice splashing and cup-size confusion in North America.

4. Solutions: forward design and system design, internal collaborative tackling, cross-domain technology migration, platformization and modularization for cost reduction, a global water-quality universal module, user validation through 1,000 questionnaires and employee-home tests, and agent factories and end-to-end factory brains to improve operational efficiency.

Summary: Platform sellers can watch how Midea manages complex businesses through platformized R&D, AI operations, and ecosystem collaboration, and how it avoids directional risks.

1. Platformized R&D practices: through the three-generation relay, the development generation is pushed toward mass production and product lineup expansion, the reserve generation lays out humanoid robots, embodied intelligence joint needs, and next-generation platforms, and the research generation targets reliability, lifespan, and forward design; industrial research is unrestricted, allowing repeated trial and error and deep problem-solving.

2. Operations management: the AI Research Institute coordinates a team of more than 400 people across Shunde, Beijing, and Shanghai, adopts agile development, and uses target products to force technology breakthroughs; AI is deeply integrated into the full R&D process, using AI to simulate material molecular structures and ionization paths and TRIZ plus AI to improve efficiency; the Jingzhou factory built the world's first agent factory, Wuxi is exploring an end-to-end factory brain, and AI agents generate hundreds of millions of yuan in annual efficiency gains.

3. Scenario operations and investment and partnership implications: industrial scenarios come first, with the Yutu inspection robot running for more than 14 days, automatically charging, and performing EHS checks, while production-line screw fastening and sorting achieve success rates above 99%; the home balcony laundry end-to-end task success rate is above 90%. Internally, the Central Research Institute, Jiya Precision, and KUKA collaborate, and a dedicated team advances industrialization of lab results, with differentiated expansion for external complete-machine customers.

4. Direction risk avoidance: maintain strategic patience, solidly build the reserve generation, do not blindly push immature products to market, and face the long journey of hardware stability and generalization.

Summary: Researchers can obtain research material from new industry directions, new problems, innovation mechanisms, and business models.

1. New industry directions: Midea achieves full-model replacement with RV reducers, with an 8,000-hour constant-load lifespan, 3,000 impacts, and 2-3 micron precision breaking records of international top competitors; micro-steam-bake surface acoustic wave technology won the 2025 National Science and Technology Progress Award; blue oxygen, plasma, and other underlying technologies are migrating to multiple categories, and the C-end technology pool is spilling over into AIDC liquid cooling, SST insulating interface materials, robot core components, and other ToB businesses.

2. New problems: core component supply chain risk and bottlenecks, mass-production consistency and quality stability, water quality differences across countries, plasma cost and reliability under high humidity, high wind speed, and dry-wet alternation, fatal consistency failures before mold opening, and humanoid robot hardware stability and generalization.

3. Mechanism and policy implications: five years and dozens of iterations without pursuing short-term output; three generations and forward design; unrestricted industrial research with no short-term revenue targets or fixed paths and no hard deadlines for deliverables; talent selection values a beautiful mind, willingness to learn, explore, and take responsibility, and resilience; and the company leads the formulation of national standards in the plasma field.

4. Business model: internal collaboration among the Central Research Institute for technology breakthroughs, Jiya Precision for manufacturing industrialization, and KUKA for complete-machine validation; differentiated expansion for external complete-machine customers; overseas OBM priority; AI agent factories and factory brains generate efficiency gains; and embodied intelligence advances with industrial use first, commercial use following, and home exploration later.