Only two companies on Earth have sustained quantum error correction across many rounds. One of them just validated an entirely different physical platform, proving that neutral atoms can run a toric code.
Artisans align polished spheres into interconnected grid patterns guided by a braided ring diagram.
Atom Computing today announced the industry’s first full demonstration of quantum error correction using a toric code on a neutral-atom system. The achievement, published June 3, 2026, places Atom Computing in a club that previously had a single member. It also vaults neutral-atom architectures from a promising science project to the leading contender for scaling fault-tolerant quantum computers. The clock on delivering a commercially relevant logical qubit just reset from a multi-year marathon to a 24-month sprint.
Why the Toric Code Matters
An agile runner in sandals outpaces armored competitors on a treacherous mountain pass toward a glowing archway.
The quantum error correction landscape has been defined by a single, brutal gate: can you run many rounds of sustained QEC and watch the logical error rate drop as you add physical qubits? One other company, using superconducting systems, cleared that bar first. Now Atom Computing has cleared it on an entirely different physical platform. The toric code is a demanding topological code that encodes information in the global pattern of a qubit lattice, not in any single physical qubit. It is the canonical test of whether a platform can scale. Atom Computing just passed it.
Prior work set the stage. In late 2024, Microsoft and Atom Computing entangled 24 logical qubits and demonstrated error detection and correction on 28 logical qubits using neutral atoms. In 2025, Infleqtion published fault-tolerant operation of logical qubits on a neutral-atom machine using a [[4,2,2]] code, showing a 12x error reduction on Bell state preparation, a 15x error reduction on random circuits, and a 6x error reduction on an Anderson Impurity Model solver. Harvard and QuEra researchers demonstrated surface code QEC with 448 neutral atoms, achieving 2.14x below-threshold performance in a four-round characterization circuit. All of this was prologue. The toric code demonstration is the main event. It proves that neutral-atom systems can execute the same class of demanding, topological error correction that makes the entire fault-tolerant roadmap credible.
How Neutral Atoms Win
The industry’s focus on gate fidelity misses the point. Atom Computing’s real weapon is connectivity. Its architecture enables dynamic rearrangement of qubits for all-to-all connectivity, removing the constraints of fixed hardware layouts found in superconducting chips. In the logical qubit era, topology is a dominant factor, and fixed-layout superconducting systems face a steep architectural challenge that neutral atoms simply bypass.
Three architectural facts from Atom Computing’s disclosure explain why this works. First, the company’s nuclear-spin qubits exhibit record-breaking coherence times. Information survives long enough to perform the many rounds of syndrome measurement that QEC demands. Second, the system’s zoned architecture supports highly parallelized operations. You can shuttle atoms between storage and processing zones, perform gates, and read out ancillas simultaneously. Third, and most critically, neutral atoms can be packed tightly in arrays held by lasers and moved around so that any qubit can interact with any other. This all-to-all connectivity is what makes the toric code’s demanding stabilizer measurements viable. Superconducting qubits are wired to their immediate neighbors. Moving logical qubits across a chip requires expensive, error-prone swap chains. Neutral atoms simply rearrange the physical qubits themselves.
“This is a historic moment for quantum computing,” said Dr. Ben Bloom, CEO and Founder of Atom Computing. “Today, we have shown that practical quantum error correction can be achieved with our neutral-atom technology.” Bloom later added: “This is the clearest demonstration yet that neutral atoms are highly competitive with superconducting systems and other approaches for building scalable logical qubits.”
The 24-Month Sprint
The Microsoft collaboration’s 24 entangled logical qubits already demonstrated that neutral atoms can build logical qubits at scale. The toric code demonstration proves those logical qubits can be protected with a top-tier error correction scheme. The combination resets the competitive timeline. The race to a commercially relevant logical qubit is no longer a multi-year marathon. It is a 24-month sprint.
Here is the prediction: within 18 months, a neutral-atom system will be the first to run a materials science simulation on logical qubits that definitively outperforms any classical supercomputer. The Infleqtion team already showed a 6x error reduction on an Anderson Impurity Model solver using logical qubits. Scaling that result to a problem size that breaks classical simulation is an engineering challenge, not a physics one. When that threshold falls, the financial services and pharmaceutical industries will move from monitoring quantum computing to budgeting for it. The technology risk is collapsing. The competitive risk of inaction is rising.
What This Means for the Reader
The error correction milestone means the path to reliable, large-scale quantum computers is no longer theoretical. Industries betting on optimization, drug discovery, and materials science now face a concrete timeline. The question has shifted from “will this work?” to “who is ready when it does?” Teams that treated quantum computing as a five-to-ten-year horizon need to compress their planning. The infrastructure for integrating quantum resources into classical HPC pipelines, the talent for writing fault-tolerant algorithms, and the budget line items for quantum access all need to move from exploration to execution.
Only two companies on Earth have sustained quantum error correction across many rounds. One of them just proved that neutral atoms are the scaling champion. The neutral-atom era of fault-tolerant quantum computing began today, and it will not wait.
