The race to build scalable, fault-tolerant quantum computers has mostly been defined by size, complexity, and enormous energy needs, until now. Canadian startup Nord Quantique is rewriting the playbook with a qubit design that integrates error correction directly into the hardware. In doing so, they may have cracked one of quantum computing’s toughest challenges: how to keep quantum machines small, stable, and energy-efficient without sacrificing performance.
Unlike traditional quantum systems that require hundreds of physical qubits to form just one logical qubit, Nord Quantique’s bosonic qubit achieves fault tolerance inside a single physical component. This radical simplification opens the door to practical, compact machines that could rival the performance of today’s most powerful supercomputers, while consuming only a fraction of the energy.
At the heart of this breakthrough is multimode encoding, a method that stores quantum information across multiple electromagnetic patterns, or “modes,” inside a superconducting aluminum cavity cooled to near absolute zero. These modes can collectively recognize and correct certain types of interference — if one mode fails, the others compensate, preserving the data’s integrity.
“The amount of physical qubits dedicated to quantum error correction has always presented a major challenge for our industry,” said Nord Quantique CEO Julien Camirand Lemyre. “Multimode encoding allows us to build quantum computers with excellent error correction capabilities, but without the impediment of all those physical qubits.”
The company’s “Tesseract code”, a type of bosonic code, adds another layer of protection by guarding against known quantum errors, such as bit flips, phase flips, control errors, and even leakage (when qubits fall out of the expected state space entirely). Leakage, notoriously hard to correct, is directly addressed in Nord Quantique’s design — a rare feat in the quantum field.
In early testing, the system successfully preserved qubit states through 32 rounds of error correction without measurable degradation, filtering out only 12.6% of test runs where the qubit didn’t behave as expected. This strong fidelity suggests a reliable platform for long-duration quantum operations, essential for real-world applications.
This technology isn’t just theoretical. Nord Quantique is planning a 100-logical-qubit machine by 2029, with a 1,000-logical-qubit system expected by 2031. Unlike today’s room-sized quantum computers requiring massive cooling infrastructure and power, their system would fit within a 215-square-foot footprint and use significantly less energy.
For context: the company estimates that breaking an 830-bit RSA encryption key would take just 1 hour and 120 kilowatt-hours on their quantum machine. A conventional supercomputer would need 9 days and 280,000 kilowatt-hours to accomplish the same.
This leap forward could have profound implications for high-performance computing (HPC) centers, where both space and energy consumption are critical.
“Beyond their smaller and more practical size, our machines will also consume a fraction of the energy,” Camirand Lemyre noted. “That makes them especially appealing to HPC centers where energy costs are top of mind.”
