Quantum computers have become a battleground for big tech corporations. Following Google and Microsoft (MS), Amazon has announced its quantum computing research findings in an international journal.
The Amazon Web Services (AWS) research team reported on the 27th in the international journal "Nature" that they have developed a quantum error correction method that is significantly more efficient than existing methods.
Traditional computers represent the absence or presence of electrons as bits (0 and 1), but quantum computers use qubits, which can exist in a superposition of 0 and 1. As the number of qubits increases, computational capability drastically improves. However, there is a critical disadvantage: qubits are extremely vulnerable to errors due to minor noise or slight vibrations. This makes quantum error correction essential.
Existing error correction methods require the addition of dozens of qubits. This is because storing information in a single qubit makes it difficult to recover when an error occurs, so information is divided among multiple qubits. This decreases computational efficiency and significantly increases hardware requirements, making the commercialization of quantum computers challenging.
The AWS research team used cat qubits for quantum error correction. The cat qubit concept is inspired by "Schrödinger's cat." Schrödinger likened a quantum state to a cat that is both alive and dead at the same time. Similarly, cat qubits can exist in two quantum states simultaneously. Conventional qubits exist in superposition of two states, such as 0 and 1, but do not exist simultaneously.
Cat qubits are very resistant to "bit flip errors" such as when 0 changes to 1 or vice versa. For instance, even if 0 turns into 1, the original 1 that existed simultaneously also changes to 0, maintaining a balance. In contrast, "phase flip errors," where the phase of the quantum state is inverted, causing the sign of calculations to change, occur frequently. Conventional superconducting qubits often face both bit flip and phase flip errors.
The research team designed an error correction code that focuses on correcting phase flip errors in cat qubits. After arranging cat qubits in superconducting quantum circuits, they successfully automated the detection and correction of errors, reducing the error rate from 1.75% to 1.65% using only five cat qubits. Traditional methods required the addition of dozens of qubits to achieve the same level of performance, but cat qubits accomplished this with significantly fewer resources.
Professor Heo Jun of Korea University's Department of Electrical and Electronic Engineering noted, "The cat qubit's inherent ability to suppress bit flip errors combined with a code to correct phase flips has successfully reduced the error rate. This hybrid method integrates the approach used by Amazon with the error correction methods of Google and IBM to implement efficient quantum error correction."
The research team stated, "This method is highly scalable, allowing it to be applicable to large-scale quantum computers, and optimizing the performance in correcting phase flip errors will bring us one step closer to fault-tolerant quantum computing (FTQC) that operates stably without errors."
References
Nature (2025), DOI: https://doi.org/10.1038/s41586-025-08642-7