Infleqtion and Nvidia have built a validated quantum error correction code that encodes 18 logical qubits in just 98 physical data qubits — roughly 5.4 physical qubits per logical one, with an 18.4 percent encoding rate. The companies announced the result at IEEE Quantum Week in Toronto this week, and the ratio matters because today's architectures often demand hundreds or thousands of physical qubits to guarantee a single reliable logical one.

The demonstration pairs Infleqtion's open-source quantum low-density parity-check (qLDPC) software library with CUDA-Q Logical, a new logical orchestration and compiler layer Nvidia unveiled at the same event. Infleqtion ran the work on an early-access version of the software.

Why the ratio is the bottleneck

Qubits are fragile. Light, sound, temperature, or the activity of neighboring qubits can strip away their quantum properties, and a machine that makes too many errors isn't commercially viable. The standard fix bundles many entangled physical qubits into one logical qubit — but the overhead is brutal.

"An unofficial law of physics is 'there's no such thing as a free lunch,'" Quantinuum executives wrote earlier this year. "Creating high quality, low error-rate logical qubits often costs many physical qubits, thus reducing the size of calculations you can run, despite your new, lower-than-ever error rates."

Infleqtion isn't alone in attacking the ratio. Sheir Yarkoni, director of quantum software for Qarakal Quantum, told The Next Platform that his startup's Pangaea architecture will cut physical qubit requirements tenfold. D-Wave has made a similar claim for its dual-rail superconducting systems in development. And Quantinuum, in a March paper, detailed technology creating 48 error-corrected and 64 error-detected logical qubits from 98 physical qubits.

What Infleqtion actually built

Victory Omole, senior quantum research engineer at Infleqtion, laid out the technical detail in a report. Using neutral atoms in the company's Sqale systems, his team used CUDA-Q Logical to build a hypergraph-product simplex (HGPS) code block, create stabilizers and logical operators, and establish the underlying algebra. The code carries 49 X checks and 49 Z checks — each detecting specific types of physical errors, each of weight six. The canonical logical representatives returned by the qLDPC software have weight four.

"Many familiar QEC constructions devote a large number of data qubits to each logical qubit," Omole wrote. "High-rate qLDPC codes change that arithmetic by encoding multiple logical qubits in one block while keeping each stabilizer check sparse. The appeal is obvious: if the rest of the architecture cooperates, more of the machine can carry useful logical information."

That conditional — "if the rest of the architecture cooperates" — is where the real work sits. Syndrome ancilla qubits, scheduling, readout, control, and classical decoding all have to preserve those headline numbers across the full system stack.

Where CUDA-Q Logical fits

CUDA-Q Logical supplies the logical ownership, block allocation, and compilation model. A small adapter converts dense NumPy arrays so the qLDPC-generated checks and logical operators can be declared as an 18-port CUDA-Q Logical code. Engineers validated the resulting block holds multiple logical qubits inside a single encoded structure, confirmed the stabilizer matrices and X/Z checks commute, and ensured two distinct logical qubit owners within one encoded block.

"This integration focuses on code construction, logical ownership, and compilation," Omole wrote. "In coming work, we plan to extend this workflow through physical syndrome extraction, neutral-atom noise modeling, detector generation, and decoder benchmarking."

The pipeline builds on architecture research by Fred Chong, Infleqtion's chief scientist for quantum software, whose lab focuses on reconfigurable neutral-atom arrays — a good fit for high-rate qLDPC codes thanks to parallel operations and flexible connectivity.

A busy year for a 2022 rebrand

Infleqtion, founded in 2007 from University of Colorado atomics research and known as ColdQuanta until its 2022 rebrand, is chasing an aggressive roadmap: more than 1,000 logical qubits by 2030. The Nvidia collaboration is the latest in a string of moves that includes a $20 million follow-on NASA contract for the Quantum Gravity Gradiometer Pathfinder, a Jet Propulsion Laboratory mission to fly the first space-based quantum gravity sensor; the opening of the Colorado Quantum Innovation Center; and three awards under the Department of Energy's Genesis Mission, which targets the world's first fault-tolerant quantum computer.

Last week, Infleqtion also announced joint R&D with Cisco to pair its neutral-atom systems and sensors with Cisco's quantum networking portfolio — an appealing match, since neutral atoms interact naturally with photons, the carriers of quantum information across networks.

Whether 5.4-to-1 survives contact with real syndrome extraction and decoder throughput will decide if Infleqtion's 1,000-logical-qubit goal for 2030 stays credible.