Nature Physics published Zhejiang University's hardware demonstration of bucket-brigade QRAM on a programmable superconducting processor, showing classical 4-bit and 8-bit data can be loaded into quantum superposition states.
A hardware-efficient gate decomposition for quantum routing nodes cut circuit depth by more than 30% versus standard controlled-SWAP designs, targeting the bucket-brigade model's O(log N) active switching.
Measured query fidelity reached 0.809±0.025 for 4-bit data but fell to 0.604±0.005 for 8-bit data, highlighting noise buildup across deeper multi-layer routing trees despite active error mitigation.
The result addresses a key data-loading bottleneck for quantum algorithms in areas such as chemical databases, fraud detection and quantum machine learning, but scaling to commercial-size memory arrays will require better gate fidelity, lower crosstalk and error correction.
As China's 'Quantum +' plan advances, is this new memory the key to winning the global technology race?
With error correction advancing so rapidly, will this QRAM breakthrough become obsolete before it can be scaled?
From Proof-of-Concept to Practicality: The 8-Bit QRAM Leap in Quantum Computing
Overview
In June 2026, researchers at Zhejiang University achieved a major milestone in quantum computing by experimentally demonstrating an 8-bit bucket-brigade Quantum Random Access Memory (QRAM). This breakthrough is important because QRAM acts as a quantum version of classical RAM, efficiently loading classical data into quantum memory. By storing data in quantum states, QRAM enables quantum algorithms to access and process large amounts of information in superposition, something not possible with current methods. This advancement directly addresses the data-loading bottleneck, paving the way for more practical and powerful quantum computers in the future.