Baidu's Integrated Superconducting Quantum Measurement and Control Platform (QianShi) – Overview and Technical Details

This article is the third report of Baidu Technology Salon Session 103, titled “Baidu Quantum Soft‑Hardware Integration”, featuring Dr. Guo Xueyi. It introduces the QianShi quantum measurement‑control platform from three aspects: the NISQ‑stage superconducting quantum measurement‑control platform, the hardware of QianShi, and the software of QianShi.

QianShi is Baidu’s industry‑grade superconducting quantum computer that integrates application, software, and hardware. Its core goal is soft‑hardware integration to provide stable, continuous quantum computing power to users, bringing quantum hardware out of the lab and making quantum computing accessible.

Historical Review

In 1962, Josephson predicted that a superconducting current could tunnel through a thin insulating layer without resistance, leading to the Josephson junction. Early experiments showed macroscopic quantum tunnelling and later, around 1985, researchers such as Clarke, Martinis, and Devoret demonstrated energy‑level quantization in Josephson‑junction currents, proving that the tunnelling current itself is quantum‑mechanical.

Since the first superconducting qubit in 1999, the coherence time of superconducting qubits has steadily increased, reaching 500 µs in 2021, an internationally leading figure.

In 2022, Baidu Quantum released QianShi, a soft‑hardware integrated superconducting quantum computer that aims to lower the barrier for quantum computing similar to classical computers.

NISQ‑Stage Superconducting Quantum Measurement‑Control Technology

During the NISQ (Noisy Intermediate‑Scale Quantum) era, three main research directions rely on a robust measurement‑control platform:

Design of multi‑qubit chips, including control line routing and inter‑qubit coupling.

Materials and process research that affect qubit coherence time.

Calibration of high‑fidelity quantum logic gates and long‑term system stability.

Hardware of the Superconducting Measurement‑Control Platform

QianShi’s platform supports up to 200‑qubit chips with 1008 measurement‑control lines. The cryogenic system provides 30 µW cooling power at 20 mK and 1000 µW at 100 mK. To reduce noise and heating, the design eliminates continuous DC bias lines, using pulsed quasi‑DC bias combined with heavily shielded microwave lines, thus avoiding the temperature rise caused by DC‑line heating.

The system employs a two‑stage cooling chain: a pulse‑tube cryocooler (~3 K) followed by a helium‑dilution refrigerator reaching millikelvin temperatures. Redundant power supplies and multiple cooling‑water loops ensure continuous operation even in case of power or water failures.

Software of the Superconducting Measurement‑Control Platform

After installing a superconducting chip, its parameters (frequency, detuning, coupler frequencies, etc.) must be characterized and quantum‑logic‑gate calibrations performed. The “LiangMai” measurement‑control software provides:

Full historical experiment record query for design feedback.

Flexible experiment creation: blank experiments with manual parameter setting or cloning of existing experiments, supporting automation.

Visualization, data plotting, and built‑in fitting programs, with the ability to add custom data‑processing scripts.

Automation scripts that, once the chip is cooled, automatically perform rapid characterization and gate calibration, reducing manual effort and operational cost.

SDK support for advanced users to script experiments (e.g., randomized benchmarking) and integrate calibrated gate pulses directly into quantum circuits.

Overall, Baidu’s superconducting quantum measurement‑control platform enables stable, long‑term operation of medium‑scale superconducting chips, offering high automation, reduced human workload, and lower maintenance costs.

Conclusion

The Baidu superconducting quantum measurement‑control platform supports the control and measurement of medium‑scale superconducting quantum chips, providing stable, automated, and cost‑effective quantum computing services.