Close up image of a quantum processsor chip

New publication in Nature

Microsoft’s application of error correction to trapped-ion qubits published in Nature

By Dr. Matthias Troyer, Technical Fellow and Corporate Vice President, Microsoft Quantum

Quantum computing is entering a new phase. Across the industry, we are seeing meaningful progress on multiple fronts at once: higher-fidelity hardware, better control systems, more practical error correction, richer software tooling, and increasingly sophisticated hybrid workflows that combine quantum, AI, and high-performance computing (HPC). Useful quantum computing will not arrive from a single breakthrough in isolation; it will come from advances that work together across the full stack.

 

Logical qubits made from trapped ions

Microsoft continues to make progress on the path to utility-scale quantum computing, as reflected in today’s publication in Nature on logical qubits. The paper, titled “Improved quantum processor logical error rates via correction and detection,” captures a set of results that marked a significant step forward for the field. By applying the Microsoft Quantum platform to Quantinuum’s trapped-ion hardware, the teams demonstrated logical error-rate improvements ranging from 11x to 800x over corresponding physical circuit baselines. In a Bell-state preparation, the logical circuit error rate was reduced from roughly 0.8% for the physical baseline to 0.001%, yielding the now well-known 800x improvement. The work also demonstrated repeated error correction with an error rate per round 51x lower than a physical baseline, along with a 22x improvement for preparing a 12-qubit cat state.

 

Microsoft quantum processor chip used in error correction research

Read the paper

“Improved quantum processor logical error rates via correction and detection,” published in Nature.

Previously, Microsoft and Quantinuum reported the most reliable logical qubits on record, involving more than 14,000 individual experiments without a single observed error and active syndrome extraction without destroying the logical qubits. We then extended that progress by creating 12 highly reliable logical qubits and demonstrating a hybrid, end-to-end chemistry simulation that combined logical qubits, AI, and HPC to estimate the ground-state energy of an important catalytic intermediate within chemical accuracy. Together, those milestones showed not only that reliable logical qubits are possible, but that they can be used in workflows that begin to solve real scientific problems. 

 

These were not isolated, single-qubit demonstrations. The paper shows fault-tolerant computation on multiple logical qubits, with circuits spanning up to 12 logical qubits. It combines two code constructions optimized for trapped-ion devices: a 12-qubit code encoding two logical qubits and a 16-qubit tesseract color code supporting larger logical computations. The result is evidence that state-of-the-art processors can use fault-tolerance techniques to strongly suppress errors in non-trivial circuits. That is a critical threshold for the entire industry.

 

For Microsoft, this work has also helped define a clear direction for the broader quantum community: The path to useful quantum computing relies on reliable logical qubits. More physical qubits alone are not enough. What matters is the ability to detect, diagnose, and correct errors so that quantum information can survive long enough to perform meaningful computation. Our qubit-virtualization approach, combined with hardware-software co-design, has helped move the discussion from whether logical qubits can outperform physical baselines to how quickly we can scale these methods across platforms.

 

Our platform works with multiple types of qubits

The Microsoft Quantum platform is an end-to-end software and systems stack designed to support multiple qubit modalities. It combines error correction, hybrid workflows, and developer tooling that can be applied across hardware architectures rather than being locked to a single approach. Our new four-dimensional geometric codes expand the field of error correction and broaden the path to reliable quantum computing. 

 

Our platform work now extends across trapped ions, neutral atoms, and our own topological qubits. With Atom Computing, we are co-developing Magne, the world’s first fault-tolerant quantum computer with 50 logical qubits. This system has been reserved by QuNorth in Copenhagen and will serve all of the Nordic countries.   

 

The Microsoft Quantum Development Kit

Developers are a crucial part of the future of quantum computing. The Microsoft Quantum Development Kit (QDK), which is part of the Microsoft Quantum platform, continues to evolve as an open-source toolkit for building quantum applications, with simulators, resource estimation, debugging tools, and a modern development experience. Integrated with Visual Studio Code and GitHub Copilot, the QDK helps developers learn, write, test, and refine quantum code faster. Today, we are also announcing “deq,” our latest open-source package for error correction as part of the QDK.

 

As the industry moves from theory to practical workflows, the ability to prototype, validate, and optimize quantum programs efficiently becomes increasingly important. We encourage developers, researchers, and students to explore the QDK, experiment with hybrid quantum applications, and build the skills that will define this next era. The QDK is what enabled the results reported above, and these tools are now available to everyone.

 

There is still much work ahead: larger codes, deeper circuits, real-time decoding, feed-forward control, and continued advances in hardware. Quantum computing is advancing because researchers are learning how to connect hardware, error correction, software, and applications into one coherent system. That is the foundation we have been building at Microsoft.  In the next few years, machines that are reliable enough to expand the frontier of chemistry, materials science, and scientific discovery will start to materialize. With our partners, our developers, and the broader research community, we are helping turn that future into something increasingly concrete—and increasingly close.

 

Learn more about Microsoft's results