China Puts a Quantum-Advantage Machine on the Cloud – and It’s Fully Domestically Built
Table of Contents
18 Dec 2025 – China Telecom Quantum Group (CTQG) and QuantumCTek have deployed the Tianyan-287, a fully domestically produced 105-qubit superconducting quantum computer, for commercial cloud access on the Tianyan quantum computing platform. The system uses the same chip architecture as USTC’s Zuchongzhi 3.0 – the processor that set the world record for quantum computational advantage in superconducting systems in March 2025.
The system’s benchmark numbers tell the story: in a random circuit sampling task using 74 qubits over 24 cycles, the Tianyan-287 generates one million samples in 18.4 minutes. The same calculation would take state-of-the-art classical supercomputers approximately 16,000 years to complete. This makes the Tianyan platform the first quantum computing cloud service outside the United States to offer quantum-advantage-class performance.
The announcement, made in late November 2025 with an accompanying arXiv paper published December 11, represents something that has not happened before in quantum computing: a laboratory research prototype demonstrating quantum advantage was re-engineered into a commercial system and deployed for public access within the same calendar year.
System specifications
The Tianyan-287 integrates 105 qubits and 182 couplers in a square grid lattice – the same architecture as Zuchongzhi 3.0. The system achieves high operational fidelities across the board:
- Single-qubit gate fidelity: 99.90%
- Two-qubit gate fidelity: 99.56%
- Readout fidelity: 98.7%
- Mean qubit T₁ relaxation time: 44.4 µs
These fidelities were measured with all qubits operating simultaneously — a more demanding test than characterizing individual qubits in isolation. The arXiv paper includes 30 consecutive days of monitoring data (August 29 to September 28, 2025) demonstrating system stability over time, a crucial metric for any commercial quantum service.
Of the 105 qubits on the chip, one is non-functional and two exhibit suppressed T₁ times due to two-level system defects – imperfections in the superconducting material that cause unwanted energy dissipation. For the 74-qubit random circuit sampling benchmark, the team selected a contiguous region of high-performing qubits and deactivated the remaining qubits by biasing them to minimum operating frequency and turning off inter-qubit couplings.
The “fully domestically produced” claim
The Global Times and other Chinese state media emphasize that the Tianyan-287 is “fully domestically produced” – a pointed claim in the context of ongoing US-China technology competition and export controls affecting quantum computing components. While the specific supply chain details are not publicly documented, the claim implies that every critical component – from the superconducting chip to the cryogenic systems, control electronics, and software stack – was manufactured within China.
This is strategically significant. Western export controls have targeted certain quantum computing components, and the ability to build a quantum-advantage-class system without imported parts demonstrates a degree of supply chain independence that few countries have achieved. Whether this extends to the most sensitive components (dilution refrigerators, specialized microwave electronics, superconducting fabrication materials) is difficult to verify from public sources.
Cqlib: The software layer
Alongside the hardware, the team released Cqlib, an open-source software development kit for interacting with the Tianyan platform. The SDK supports the full workflow from circuit construction through transpilation, execution, and result analysis. Notably, it includes adapters for compatibility with Qiskit, Cirq, and PennyLane – the three dominant Western quantum software frameworks.
This interoperability is a pragmatic choice. By making Tianyan accessible through familiar programming interfaces, CTQG lowers the barrier for international researchers and developers who want to run experiments on Chinese quantum hardware without learning a new toolchain. It also signals that China’s quantum cloud ambitions extend beyond domestic use.
The Tianyan platform as of November 2025
The Tianyan-287 joins an expanding fleet. As of mid-2025, the Tianyan quantum computing cloud hosts five quantum systems: the Tianyan-287 (105 qubits), the Tianyan-504 (504 qubits), two 176-qubit machines, and a 24-qubit system – totaling 880 superconducting qubits across the cluster. Since the platform’s launch in November 2023, it has received more than 37 million visits from users across 60 countries and processed over 2 million experimental tasks.
Comparison with Zuchongzhi 3.0
The Tianyan-287’s quantum advantage claim is more modest than Zuchongzhi 3.0’s laboratory result. Zuchongzhi 3.0, using 83 qubits over 32 cycles, achieved a 10¹⁵-fold speedup over classical supercomputers. The Tianyan-287 benchmark uses fewer qubits (74) and fewer cycles (24), yielding a smaller but still decisive advantage – roughly 450 million times (approximately 10⁸·⁷) faster than classical computation, with the 16,000-year classical estimate.
This gap is expected. A commercial cloud system optimized for stability, uptime, and user accessibility will not match the performance of a laboratory prototype where researchers can spend weeks calibrating a single experiment. The fact that quantum advantage persists even in commercial operating conditions – with all the constraints of continuous uptime, automatic calibration, and multi-user access – is itself a meaningful result.
From Lab Record to Cloud Service in Nine Months
I have covered every major Zuchongzhi milestone on this site, and each time the story was fundamentally the same: a laboratory demonstration that pushed the frontier of what superconducting quantum hardware could do. The Tianyan-287 is a different kind of story. This is not a laboratory breakthrough. This is a product launch.
And that should change how we think about China’s quantum program.
The lab-to-cloud gap, closed
In quantum computing, there has always been a sharp divide between what gets demonstrated in a research lab and what gets deployed as a service. Google’s Sycamore achieved quantum supremacy in 2019, but Google has never offered a Sycamore-class machine for public cloud access. IBM offers its processors via the cloud, but its most powerful public systems are not the same chips that produce their headline research results. There is typically a multi-year lag between a research demonstration and a commercially accessible system built on the same technology.
China just compressed that lag to nine months.
In March 2025, Zuchongzhi 3.0 set the world record for quantum computational advantage in superconducting systems. By November 2025, a system using the same chip architecture is deployed as a commercial cloud service, with documented 30-day stability, an open-source SDK, and compatibility with Western software frameworks. The quantum advantage is smaller (10⁸·⁷ vs. 10¹⁵), because the commercial benchmark uses fewer qubits and shallower circuits to prioritize reliability over peak performance. But it’s still quantum advantage – on a commercial system, available to anyone.
No other country has done this.
The 880-qubit cluster: heterogeneous quantum computing
What I find most strategically interesting about the Tianyan platform is not any single machine – it’s the portfolio. Five quantum systems totaling 880 qubits, ranging from 24 to 504 qubits, all accessible through a unified cloud interface. This is heterogeneous quantum computing – different machines optimized for different tasks, orchestrated through a common software layer.
This approach mirrors how classical high-performance computing evolved. No one runs all their workloads on a single supercomputer. Different problems call for different architectures – GPUs for parallel computation, CPUs for serial tasks, FPGAs for specialized processing. The Tianyan platform is building the quantum equivalent: a fleet of processors with different qubit counts, fidelity profiles, and connectivity patterns, available to researchers who can choose the right system for their experiment.
IBM has been moving in this direction with its Quantum Network and multi-chip architectures. But the Tianyan platform has an advantage: it’s backed by a state telecommunications company (China Telecom) with existing data center infrastructure, global network connectivity, and a mandate to make quantum computing commercially available. This is not a startup burning through VC funding while hoping to find product-market fit. This is a national champion deploying quantum infrastructure as a service.
What “fully domestically produced” means
The emphasis on domestic production is not just rhetoric. It reflects a strategic calculation about supply chain resilience that maps directly onto the broader US-China technology competition.
Quantum computing systems depend on a small number of specialized components: dilution refrigerators (which cool processors to ~15 millikelvins), arbitrary waveform generators (which produce the microwave pulses that control qubits), high-frequency coaxial cables, and superconducting chip fabrication tools. Historically, several of these components were sourced from Western suppliers – particularly Finnish company Bluefors for dilution refrigerators and various US and European companies for microwave electronics.
If the “fully domestically produced” claim is accurate, it means China has developed or sourced domestic alternatives for all of these components. Given that the US and allied governments have been tightening export controls on quantum-related technologies, this supply chain independence is not a luxury – it’s a strategic necessity for any country that wants to sustain a sovereign quantum computing program.
For Western policymakers, this is a sobering data point. Export controls on quantum components were designed, in part, to slow China’s progress. If China can build quantum-advantage-class systems without imported components, those controls may be less effective than assumed.
Quantum Upside & Quantum Risk - Handled
My company - Applied Quantum - helps governments, enterprises, and investors prepare for both the upside and the risk of quantum technologies. We deliver concise board and investor briefings; demystify quantum computing, sensing, and communications; craft national and corporate strategies to capture advantage; and turn plans into delivery. We help you mitigate the quantum risk by executing crypto‑inventory, crypto‑agility implementation, PQC migration, and broader defenses against the quantum threat. We run vendor due diligence, proof‑of‑value pilots, standards and policy alignment, workforce training, and procurement support, then oversee implementation across your organization. Contact me if you want help.
Related Content
