Amazon Web Services (AWS) has launched its new quantum computing chip, “Ocelot.” This move puts AWS in competition with Microsoft and Google. AWS follows Google’s December 2024 announcement of the “Willow” chip and Microsoft’s introduction of “Majorana 1” in February 2025. According to Amazon’s research team, Ocelot leverages an innovative approach inspired by traditional chip transistors, promising a more efficient utilization of qubits—the fundamental units of quantum computing.
In a recent research paper published in Nature, lead author Harald Putterman from the AWS Center for Quantum Computing in Pasadena, California, detailed how analog circuits, rather than conventional digital 1s and 0s, could significantly reduce the number of physical qubits required for quantum computing. AWS expanded on these findings in a blog post, explaining the concept in a more accessible manner.
Quantum computing faces a significant challenge: consolidating enough physical qubits to counteract errors and form stable “logical qubits” capable of reliable mathematical operations. Each company has chosen a different approach to this issue. Google employs standard superconducting materials in its Willow chip, while Microsoft’s Majorana 1 utilizes rare Majorana particles- unique particles whose antiparticles are identical. AWS, however, is betting on “cat qubits,” an approach they believe could surpass existing methods in efficiency and reliability.
こちらもお読みください： NTTとドコモ、ネットワークとサービスの統合によるコンピューティングサービスのオンデマンド統合制御の実証に成功
According to AWS, “Ocelot is our first chip based on the cat qubit architecture, serving as an initial test of its viability for quantum error correction.” The term “cat qubit” is derived from Schrödinger’s cat, a famous thought experiment in quantum mechanics that explores the concept of superposition- where a cat is both alive and dead until observed. Unlike digital qubits, which are measured as discrete 1s and 0s, cat qubits are analog and exist as continuous wave values. In this case, the wave amplitudes represent a collection of trapped photons in a light-shaping waveguide.
The chip’s name, Ocelot, is a playful nod to both wild cats native to the Americas and the “cat qubits” that form its foundation. Analog computing in quantum mechanics has been studied for over two decades, with pioneers like John Preskill of UCLA exploring its potential to enhance qubit efficiency. AWS’s research builds upon experiments conducted in 2019 by the French National Institute for Digital Science and Technology (INRIA).
Putterman’s team claims that their cat qubit measurements allow them to achieve quantum error correction with significantly fewer qubits- five data qubits and four auxiliary qubits—compared to the 49 qubits required in traditional surface code devices. AWS likens this efficiency gain to the transformative impact of silicon transistors on classical computing. “History has shown that scaling the right components can dramatically affect cost, performance, and feasibility,” AWS stated in its blog.
The shift from digital to analog computing aligns with a broader movement toward harnessing the advantages of analog electronics. Unlike digital systems, which count discrete values, analog computing measures continuous variables, enabling more flexible data manipulation. Despite current limitations- such as Ocelot’s five-qubit capacity, which is insufficient for logical operations- AWS researchers remain optimistic about the technology’s potential.
Like their counterparts at Google and Microsoft, AWS scientists believe they have chosen the right qubit architecture to scale quantum computing effectively. “Ocelot’s hardware-efficient approach to error correction positions it well for the next phase of quantum computing: scaling. By adopting this strategy, we can accelerate the development of cost-effective, error-correcting quantum computers that will ultimately benefit society,” AWS concluded in its blog post.
