新型自旋電晶體設計有望提升 AI 運算效能
New spin transistor design could boost AI computing
Updated at: July 22, 2026 at 02:15 AM
隨著人工智慧持續進步，資料中心的能源需求已達臨界點。
As artificial intelligence continues to advance, the energy demands of data centers are reaching a breaking point.
目前，電腦受限於「馮紐曼瓶頸」，即在分開的記憶體與處理單元間傳輸資料會造成能源浪費。
Currently, computers suffer from the 'Von Neumann bottleneck,' where energy is wasted moving data between separate memory and processing units.
波士頓學院的研究人員揭露了一種使用名為CrSBr獨特材料的新型電晶體設計。
Researchers from Boston College have unveiled a new transistor design using a unique material called CrSBr.
這種雙功能材料允許運算與儲存整合，顯著提升了效率。
This dual-function material allows for integrated computing and storage, significantly increasing efficiency.
透過利用電子的「自旋」而非僅僅是移動電荷，此元件所需電力大幅減少。
By utilizing the 'spin' of electrons rather than just moving electrical charges, this device requires far less power.
此外，由於這些基於自旋的裝置具有非揮發性，它們即使在沒有持續供電的情況下也能保留資料，消除了待機時的電力浪費。
Furthermore, because these spin-based devices are non-volatile, they retain data even without constant electricity, eliminating standby power waste.
此項創新與標準製造製程相容，標誌著AI硬體邁向更永續與高效未來的重大一步，將典範從單純基於電荷的系統轉移至整合式基於自旋的技術。
Compatible with standard manufacturing processes, this innovation marks a major step toward a more sustainable and efficient future for AI hardware, shifting the paradigm from purely charge-based systems to integrated spin-based technology.
