Stability First: Amazon Secures Patent for Scalable Quantum Systems with Fluxonium Qubits

Quantum computers are fascinating technology, but they still have their quirks: the main challenge lies in the stability of qubits, the quantum bits that are fundamental to their operation. Amazon, which has already made an impression in this field with its Ocelot chip (which is said to reduce error correction costs by up to 90%), has filed a new patent. The application is titled «Reading and Resetting Fluxonium Qubits.»

It might sound like something out of fantasy, but fluxonium qubits are very much a real type of qubit. Their main advantages include greater stability and longer coherence times compared to some other popular qubits. This means that these qubits can maintain their delicate quantum state for a longer period, which is crucial for complex computations.

Amazon’s system aims to carefully read the state of a fluxonium qubit after an operation and reliably reset it to its initial state, preparing it for the next task. How does it achieve this? It employs a specialized **resonator for reading**, which extracts information from the qubit and transmits it to a **quantum metamaterial**—a material designed with exotic properties specifically for interacting with signals from quantum components. The key feature here is that this process allows the qubit’s state to be determined without destroying it or introducing unnecessary noise.

The goal of all this is to make quantum algorithms **more accurate, faster, and efficient**. Patents of this caliber signal that giants like Amazon are serious about overcoming the main barriers to scalable and practical quantum computing.

Although Amazon hasn’t launched a fully operational quantum computer yet, it has long been involved in the field through its **Amazon Braket** service within AWS, which provides remote access to quantum devices from various manufacturers. The company acts both as a platform and as a hardware developer (with the Ocelot chip) and now also as an innovator in fundamental qubit management techniques.

**To deploy AI models—from Claude Opus 4 and DeepSeek R1–0528 to Midjourney 7 and Flux 1.1 Pro Ultra—anyone can do it. Just check out BotHub: it’s a convenient aggregator where all the top neural networks are collected in one place.** **Register via this link** **and get a bonus of 100,000 tokens.**