Full-stack Analog Optical Quantum Computer with A Hundred Inputs
Shota Yokoyama, Atsushi Sakaguchi, Warit Asavanant, Kan Takase, Yi-Ru Chen, Hironari Nagayoshi, Jun-ichi Yoshikawa, Takahiro Kashiwazaki, Asuka Inoue, Takeshi Umeki, Toshikazu Hashimoto, Takuji Hiraoka, Akira Furusawa, Hidehiro Yonezawa · 2025
Read Full-stack Analog Optical Quantum Computer with A Hundred Inputs by Shota Yokoyama, Atsushi Sakaguchi, Warit Asavanant, Kan Takase, Yi-Ru Chen, Hironari Nagayoshi, Jun-ichi Yoshikawa, Takahiro Kashiwazaki, Asuka Inoue, Takeshi Umeki, Toshikazu Hashimoto, Takuji Hiraoka, Akira Furusawa, Hidehiro Yonezawa (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-16147.txt.
Introduction
Shota Yokoyama and co-authors present a quantum computing paper, published in 2025, reporting a full optical quantum platform with one hundred inputs. Light based machines promise huge, ultrafast computation, yet making one programmable at scale has been the hard part. Their system runs at a clock speed of one hundred megahertz and is steered through a cloud interface with an open Python toolkit.
The team characterizes the machine through multi input, multi step teleportation of quantum states and programmable routing across one hundred one input modes. They frame the platform as a working testbed for the next steps: adding the non Gaussian resources that full quantum advantage needs and growing optical neural networks.
Catalog record
| CREATOR | Shota Yokoyama, Atsushi Sakaguchi, Warit Asavanant, Kan Takase, Yi-Ru Chen, Hironari Nagayoshi, Jun-ichi Yoshikawa, Takahiro Kashiwazaki, Asuka Inoue, Takeshi Umeki, Toshikazu Hashimoto, Takuji Hiraoka, Akira Furusawa, Hidehiro Yonezawa |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 71 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-16147
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Full-stack Analog Optical Quantum Computer with A Hundred Inputs Shota Yokoyama, Atsushi Sakaguchi, Warit Asavanant, Kan Takase, Yi-Ru Chen, Hironari Nagayoshi, Jun-ichi Yoshikawa, Takahiro Kashiwazaki, Asuka Inoue, Takeshi Umeki, Toshikazu Hashimoto, Takuji Hiraoka, Akira Furusawa, Hidehiro Yonezawa Abstract † † equal-contributors: These authors contributed equally to this work. † † equal-contributors: These authors contributed equally to this work. Optical technology emerges as a highly promising platform for quantum computing, driven by its enormous potential for large-scale ultrafast computation and its integration with telecom technology asavanant2022optical ; takeda2019toward ; o2007optical . There have been intensive investigations ongoing into the development of optical quantum computers, however, they are limited to small-scale, special-purpose, or low-quality systems zhong2020quantum ; asavanant2021time ; larsen2021deterministic ; aghaee2025scaling . Error correction for fault tolerance is still very challenging, making a full-scale fault-tolerant quantum computer a long-term goal. However, practical testbeds for quantum computers, without the difficulty in error correction, are in high demand. Here we present an innovative analog optical quantum computer utilising continuous variables. Our analog optical quantum computer is based on sequential measurements of time-domain-multiplexed large-scale two-dimensional entanglement asavanant2019generation ; larsen2019deterministic . Accumulated Gaussian noise and spurious bias caused by the imperfection of analog computing can be suppressed through careful calibration and repeated trials without complicated error correction. Our system achieves a hundred analog inputs and operates at a clock frequency of 100\text{\,}\mathrm{MHz} with a comprehensive full-stack architecture featuring a cloud interface and a Python software development kit (SDK) for enhanced accessibility and scalability. We demonstrate the detailed characterisation of our optical quantum computer and quantum state sorting as its example application. This development marks a significant step forward in the exploration of analog quantum computing, with a potential to accelerate both fundamental research and practical applications such as fast and large-scale optical neural networks killoran2019continuous ; bangar2023experimentally ; anand2024time ; bangar2025continuous . Quantum computer is believed to be a game changer in a field of information technology. Since a quantum computer works by quantum mechanics, the different principle from that of the modern computer, it can efficiently solve certain problems that the modern computer cannot solve as efficiently nielsen2010quantum ; gyongyosi2019survey ; gill2022quantum . Quantum computer has a strong potential for applications in material science bauer2020quantum , finance herman2023quantum , machine learning perdomo2018opportunities and so on. Quantum computers have been developed with a variety of physical platforms ladd2010quantum including superconducting circuits, neutral atoms, ions, silicons, and optics. While quantum error correction has been intensively investigated, current quantum computers are still limited to noisy intermediate-scale quantum (NISQ) computer or barely claim quantum supremacy with around a few hundred qubits without any practical industrial applications preskill2018quantum ; arute2019quantum ; zhu2022quantum ; kim2023evidence ; ai2024quantum ; gao2025establishing . A full-fledged fault-tolerant quantum computer is still a long-term goal. While most quantum computers are based on qubits and are considered digital computers, the alternative computing approach with analog variables, i.e., analog computing, is worth revisiting for developing practically useful quantum computers. Traditionally, an analog computer features low energy consumption and fast processing, but limited precision, or relatively high cost to improve precision maclennan2009analog ; maclennan2014promise . Analog computing is particularly suitable for nature-inspired computational models such as neural networks krotov2023new ; mead1988silicon . The state of a neuron is represented by continuous variables (CVs), which makes it natural to implement using analog computer. Neural network often does not require high precision as neural information processed in the brain actually operates with very low precisions maclennan2009analog . Thus, quantum computing with continuous variables braunstein2005quantum has the potential to serve an alternative approach for developing a practically useful near-term quantum computer for such tasks. Like classical analog computing, analog quantum computing is considered fast at producing approximate solutions and suitable for problems based on continuous variables, such as neural networks killoran2019continuous ; bangar2023experimentally ; anand2024time ; bangar2025continuous and quantum simulation kendon2010quantum . Demonstrations of analog quantum computing with continuous variables have been reported in an optical domain asavanant2021time ; larsen2021deterministic . Optical systems are ideal platforms for analog quantum computing because CV quantum resource states and efficient measurements are readily available, and measurement-based computation raussendorf2001one ; menicucci2006universal ; briegel2009measurement ; menicucci2011temporal with time-domain-multiplexing technique yokoyama2013ultra ; yoshikawa2016invited ; asavanant2019generation ; larsen2019deterministic makes it possible to perform a large-scale quantum computation with a compact setup. The previous demonstrations asavanant2021time ; larsen2021deterministic , however, were not practically useful due to their limited number of inputs and low clock frequency which is the frequency of measurements and represents the speed of a basic operation. The demonstrations were limited to single input mode at 25\text{\,}\mathrm{MHz} clock frequency asavanant2021time , and six input modes at 4\text{\,}\mathrm{MHz} clock frequency larsen2021deterministic , respectively, as optical parametric oscillators in their systems inherently limited the overall bandwidth as well as the number of inputs. Demonstrations of small-scale systems do not necessarily guarantee large-scale execution, since accumulated noise and spurious bias may render the outputs meaningless after a certain number of computational steps. In addition, a flexible design platform for quantum circuits was not available, which made it hard to construct complex and large-scale quantum circuits. In this article, we demonstrate a measurement-based optical quantum computer operating at 100\text{\,}\mathrm{MHz} clock frequency with a hundred input modes, using ultra-wideband optical parametric amplifiers (OPAs) kashiwazaki2023over . Here, the clock frequency corresponds to the implementation speed of a two-mode operation in our system. Our OPAs feature 6\text{\,}\mathrm{THz} bandwidth, which unlock the potential of an optical quantum computer. Although the current 100\text{\,}\mathrm{MHz} clock frequency already surpasses the gate speeds of conventional platforms, achieving THz-clock quantum computing may be feasible in the future yamashima2025all . Our optical quantum computer utilises a time-domain-multiplexed quad-rail lattice cluster state alexander2016flexible . Quantum operations are implemented by measuring the entanglement, followed by feedforward operations. By properly designing the measurement bases and feedforward gains, we can implement arbitrary multi-mode Gaussian operations for multiple inputs. Accumulated Gaussian noise and spurious bias through multiple computational steps are suppressed by careful calibration and repeated trials without the need for a specific error correction. Furthermore, a cloud-based system with the tailored Python SDK is built. Our quantum computer is accessible anywhere through the cloud, and flexible quantum circuit design is easily executed by the SDK as in qubit-based quantum computers gill2022quantum . These systems make our computer a practical testbed for investigating applications of an analog quantum computer. It is also worth noting that our optical quantum computer is not just an analog quantum computer designed for near-term applications, but a solid platform for a future full-stack fault-tolerant quantum computer. Recently, generations of high-quality optical Gottesman–Kitaev–Preskill gottesman2001encoding (GKP) qubits, which are promising qubit encodings for error correction, have been reported konno2024logical ; larsen2025integrated . On our optical quantum computer, basic operations on GKP qubits, including entangling operations and error syndrome measurements, can be implemented, laying the foundation for a universal fault-tolerant quantum computer konno2024logical ; walshe2021streamlined . We evaluate single-mode and two-mode operations, verifying the quantumness of our quantum computer as well as demonstrating multi-input, multi-step operations. In particular, we demonstrate 101-modes 100-step parallel quantum teleportation, evidencing that unwanted noise and bias introduced through multiple operations can be suppressed and eliminated. In addition, we demonstrate quantum state sorting by their amplitudes, where we apply random