After the publication of a scientific paper about Amazon’s breakthrough in quantum computing, it’s time to learn a little more about the quantum processor for this platform. It’s a prototype called “Ocelot,” which echoes the name of Amazon, since these feline predators live along the South American river of the same name. And cats are appropriate here, since the processor’s architecture is based on cat qubits.
Image source: Amazon
We discussed the basics of the quantum platform with the Ocelot processor in detail in this news item. Let us recall that Amazon engineers combined two different qubit architectures in one processor, thereby achieving a radical reduction in hardware costs — the number of physical qubits required to implement error correction schemes in calculations.
According to the company, Ocelot’s circuits correct errors with 90% less physical qubits than competing platforms. In other words, Amazon’s breakthrough quantum processor uses an order of magnitude less hardware resources to accurately execute quantum algorithms. Needless to say, this has huge potential for more easily scaling up the number of qubits.
Image source: Nature 2025
Amazon describes the design of the processor this way: “The Ocelot memory logic chip, shown in the diagram above, consists of five cat-qubit data units, each of which contains an oscillator used to store quantum data. Each cat-qubit’s reference oscillator is connected to two auxiliary transmon qubits for phase-shift error detection, and is interfaced to a special nonlinear buffer circuit used to stabilize the cat-qubit states and exponentially suppress bit-shift errors.”
Cat qubits, named after Schrödinger’s fictional cat (the original is a cat, not a tomcat), are resistant to bit-flip errors because they use groups of photons and neglect flips of one of them. Transmons are used to correct phase-shift errors and correct the only errors that cat qubits make. This hybrid architecture eliminates errors in quantum computing in a simpler way.
«Tuning the Ocelot device involves calibrating the bit and phase error rates of the cat qubits as a function of the cat amplitude (the average number of photons) and optimizing the noise offset of the C-NOT gate used to detect phase errors. Our experimental results show that we can achieve bit switching times approaching one second, which is more than a thousand times longer than the lifetime of conventional superconducting qubits,” Amazon explains.
Physically, the Ocelot chip consists of two electrically connected crystals, each 1 cm2 in area. On the surface of each silicon microchip are thin layers of superconducting materials that form the elements of quantum circuits. The Ocelot chip consists of 14 main components: five data qubits (cat qubits), five “buffer circuits” to stabilize the data qubits, and four additional qubits to detect errors in the data qubits (transmons).
Quantum bits store the quantum states used for computation. To do this, they rely on components called oscillators, which generate a repeating electrical signal at a constant frequency. Ocelot’s high-quality oscillators are made from a thin film of a superconducting material called tantalum. AWS’s materials scientists developed a special way to process tantalum on a silicon chip to improve the oscillator’s performance.
Overall, the company has borrowed most of the technology for Ocelot from the semiconductor industry and is poised to quickly cut the cost of producing processors by a factor of five. Amazon’s cutting-edge quantum platform — fundamentally different from all the others — should speed up the emergence of a practical and error-resistant quantum computer by five years, the company is confident.