Anyon Computing has introduced a groundbreaking open quantum control system that integrates quantum processors as co-processor nodes alongside graphical processing units (GPUs) and central processing units (CPUs) within a single apparatus.
This innovative design diverges from conventional “quantum box” configurations. Relying on NVIDIA’s NVQLink, the system facilitates the measurement of quantum-classical loops within microseconds, an essential capability for preserving qubit coherence and addressing computational challenges that surpass the limitations of individual chips.
“A quantum supercomputer must function as a unified machine, rather than a quantum box affixed to a classical counterpart,” asserts Roger Luo, Co-Founder and CEO of Anyon Computing, elucidating the design philosophy underpinning this architecture, which is intended for the forthcoming evolution of data centers.
Anyon’s NVQLink Integration Fosters Quantum-Classical Co-Processing
The newly launched quantum control system from Anyon Computing leverages NVIDIA NVQLink to attain microsecond-level measurements of quantum-classical interactions, a speed crucial for sustaining qubit coherence throughout intricate computations.
This architecture marks a significant departure from traditional configurations, wherein quantum processors operate as isolated components tethered to classical systems; instead, they function as integral components of a larger, harmonized machine.
This paradigm effectively addresses a pivotal issue in scaling quantum computing: the exigent need for expedited feedback between quantum measurements and classical processing.
Calibration, control, and error correction demand computational resources that cannot be confined to a single chip, prompting a need for a tightly integrated system that minimizes data transfer latency.
By operating NVIDIA CUDA-Q programs natively, Anyon’s framework enables adaptive quantum-classical loops to function at rhythmical cadences ranging from milliseconds to microseconds, avoiding delays inherent in recompilation processes.
This enhanced speed is supported by the simultaneous employment of commodity RDMA-over-Ethernet links for communication between host nodes and peer controllers, which empowers the control plane to expand from a solitary node to an extensive, interconnected network.
The development of this avant-garde system was propelled by autonomous AI agents, representing a novel methodology for constructing intricate quantum computing frameworks.
According to Anyon, the integration of its system with NVIDIA’s infrastructure is a pivotal facilitator of progress.
“The creation of functional quantum-GPU supercomputers presents a challenge of heterogeneous systems, as the efficiency of quantum processors relies on tightly coupled access to accelerated computing resources,” explains Sam Stanwyck, Director of Quantum Product at NVIDIA.
“Anyon Computing has demonstrated that utilizing NVIDIA NVQLink provides the essential environment for the most critical workloads in large-scale quantum computing, encompassing calibration, control, and error correction.”
The system’s architecture fosters scalability, with controllers phase-locked to a shared reference, reflecting conventional data center growth patterns and laying the groundwork for increasingly powerful quantum-classical hybrid systems.
AI-Driven Development of Real-Time Quantum Control Systems
This architecture fundamentally transforms the interaction between quantum and classical resources, transcending the traditional view of quantum processors as mere peripheral components linked to host computers.
The immense computational demands associated with calibration, control, and particularly error correction surpass those manageable by a single processor.
These AI agents autonomously optimized the design, evolving from initial register-transfer-level coding through simulation, construction, and hardware-in-the-loop validation, continuously refining system characteristics with minimal human oversight.
This represents one of the pioneering instances of an advanced, real-time FPGA system culminating in production release via such an automated procedure.
“We have modeled our development approach on this blueprint for the future, with AI agents performing engineering tasks against actual hardware,” states Anyon Computing.
In November 2025, NVIDIA expanded the NVQLink open reference architecture in collaboration with Oxford Quantum Circuits and Orca Computing, further consolidating its role as a key enabler of hybrid quantum-classical computing, according to the firm.
The recent introduction of CUDA-Q logical, augmented by quantum error correction, exhibits a sustained commitment to constructing fault-tolerant quantum systems.
Anyon’s framework is poised to be open-sourced, establishing it as a foundational control plane for emerging quantum data centers and expediting the progression of scalable, real-time quantum computing.
Source link: Quantumzeitgeist.com.
