Working Principle: Embodied Intelligence with Self-Sensing and Self-Actuation
Shape Memory Alloys (primarily Nickel-Titanium or NiTi) exhibit a unique "Shape Memory Effect." The material undergoes plastic deformation at low temperatures. However, once the temperature rises to a specific phase transformation point (e.g., 60–80°C), an internal martensitic phase transformation occurs. This transition releases massive recovery stress, causing the component to revert to its preset shape with high precision and force.
By utilizing a custom nitinol spring or niti spring, the entire process requires no external power supply or signal input. This achieves a hardware-level fusion of temperature "sensing" and mechanical "actuation."
Applications: Achieving a Revolution in Ultra-Simplified Architecture
In server liquid cooling, this technology leverages structural innovation to solve core thermal management bottlenecks:
- Self-Actuated Thermal Control Valves: By integrating an sma spring or specialized alloy valve plates directly into the liquid cooling loop, the alloy deforms automatically if the coolant temperature exceeds thresholds. This deformation directly drives the valve core to regulate flow rates or hot/cold mixing ratios in real time, enabling completely passive, closed-loop control.
- Key Components in Immersion Cooling: Within immersion cooling tanks, this shape memory behavior functions as a reliable nitinol actuator to drive sealing mechanisms or bubble-elimination components. This effectively mitigates leakage risks and gas accumulation during server hot-swapping, significantly boosting system reliability.
- System Value: Using a single intelligent material component to replace independent temperature sensors, signal cables, control chips, and drive circuits drastically simplifies system architecture. It reduces material costs (BOM), eliminates power consumption, and minimizes potential points of failure.
Industry Progress: China Leads Commercialization through Integrated Manufacturing
Domestic Market (Advanced Application Stage):
- Ruijie Networks has filed patents for "shape memory alloy valve plates," achieving device-level, passive self-regulation of coolant flow.
- Qiangrui Precision has integrated a "double-seal + shape memory alloy" structure into its liquid cooling quick-disconnect couplings (compatible with Nvidia ecosystems) to ensure reliable hot-swap sealing, which is already in mass production.
- Patents from companies like Huawei further validate the feasibility of replacing electronic sensors in cold-plate liquid cooling systems.
- Core Characteristics: Domestic innovation focuses on component-level structural simplification and cost control. This aligns perfectly with high-end manufacturing and rapid commercialization needs, establishing clear product and patent barriers.
Overseas Market (Focus on Frontier Exploration):
- Early patents (e.g., Google, IBM) mostly involved using SMA for physical adjustments in air-cooling architectures, which did not become mainstream in liquid cooling.
- Current academic and R&D focus centers on more ambitious "solid-state cooling" systems. For instance, the Hong Kong University of Science and Technology (HKUST) published research in Nature regarding a kilowatt-scale elastocaloric cooling device. It utilizes the stress-induced phase transformation cycle of SMA for refrigeration. While seen as a disruptive green cooling technology, it remains in the laboratory prototype stage.
- Core Difference: Overseas research prioritizes system-level energy efficiency revolutions and cutting-edge science over component-level replacement and manufacturing integration for existing liquid cooling setups.
Industry & Investment Insights
- Key Beneficiaries: This technology directly benefits shape memory alloy material suppliers and core structural component manufacturers with precision machining capabilities. The value growth is driven by functional materials replacing traditional electronic components.
- Critical Path to Liquid Cooling Cost Reduction: This innovation represents a crucial "micro-innovation" that drives the large-scale commercialization of liquid cooling solutions and lowers per-unit thermal management costs. It will accelerate the penetration of liquid cooling in AI computing infrastructure.
- Technology Diffusion Trend: This shift confirms the current tech investment rotation, moving from digital computing power (chips/optical modules) to physical infrastructure support (power/energy, high-end manufacturing, and advanced materials). It underscores the immense value of "manufacturing barriers" and "materials innovation" within hard tech.