量子霸權爭奪戰升溫：中國「祖沖之3.2」打破美國壟斷，成全球第二個跨越關鍵門檻者
中國量子電腦發展迎來重大突破。「祖沖之3.2」成為繼Google之後，全球第二個達到「容錯閾值」的量子電腦，打破美國在此領域的壟斷，預示美中科技競爭將更趨激烈。
全球量子電腦競賽的格局正在改變。中國研究團隊在打造實用級量子電腦的競賽中邁出關鍵一步，成為美國以外第一個、全球繼Google之後第二個跨越「容錯閾值」的團隊。這項成就意味著，量子電腦大規模可靠運作的可能性已不再是美國的專利。
「祖沖之3.2」的里程碑
由中國科學技術大學學者潘建偉領導的團隊宣布，其開發的超導量子電腦「祖沖之3.2」已達到容錯閾值（fault-tolerant threshold）。這是一個舉足輕重的技術指標，它決定了機器在修正計算錯誤方面的能力，是通往大規模、高穩定性量子計算的必經之路。
何謂容錯閾值？：量子位元極不穩定，容易產生錯誤。當錯誤修正能力超過錯誤發生率時，就跨過了此門檻。這代表理論上可以建構出能持續進行複雜運算、不會因錯誤累積而崩潰的量子電腦。
美中科技競賽白熱化
此前，全球僅有Google公開宣稱達到此一成就。中國團隊的突破，不僅是科學上的勝利，更在地緣政治上釋放出強烈訊號：在足以顛覆國安、金融與製藥等產業的量子領域，美中之間的短兵相接將會更加激烈。
This content is AI-generated based on source articles. While we strive for accuracy, errors may occur. We recommend verifying with the original source.
Related Articles
Chinese researchers have achieved a critical milestone in quantum computing with their Zuchongzhi 3.2 processor, reaching the fault-tolerant threshold. They are now the second team globally, after Google, to clear this hurdle, intensifying the tech race.
VirusTotal founder Bernardo Quintero spent decades searching for the anonymous creator of a 1992 virus that launched his career. The bittersweet resolution reveals a story of unintended consequences that transformed a Spanish city into a tech hub.
More AI agents isn't always better. A joint study from Google and MIT provides a quantitative answer to the optimal size and structure of AI agent systems, with key guidelines for developers and decision-makers.
AI coding agents from OpenAI, Anthropic, and Google are transforming software development. Understand how LLM technology works, its potential pitfalls, and what developers need to know.
Chinese researchers have achieved a critical milestone in quantum computing with their Zuchongzhi 3.2 processor, reaching the fault-tolerant threshold. They are now the second team globally, after Google, to clear this hurdle, intensifying the tech race.
VirusTotal founder Bernardo Quintero spent decades searching for the anonymous creator of a 1992 virus that launched his career. The bittersweet resolution reveals a story of unintended consequences that transformed a Spanish city into a tech hub.
More AI agents isn't always better. A joint study from Google and MIT provides a quantitative answer to the optimal size and structure of AI agent systems, with key guidelines for developers and decision-makers.
AI coding agents from OpenAI, Anthropic, and Google are transforming software development. Understand how LLM technology works, its potential pitfalls, and what developers need to know.