Innovation Summer 2026

“

• Quantum gate-based computing (circuit model): This method mimics classical logic gates using quantum bits, allowing complex algorithm execution by chaining reversible quantum logic gates. It is commercialized and widely researched and is theoretically capable of solving any problem a classical computer can; in some commercially and scientifically relevant cases, it can do so more efficiently. • Quantum annealing : Unlike the circuit model, quantum annealing optimizes restricted problem sets by leveraging quantum tunnelling. It rapidly finds near-optimal solutions for NP-hard problems, making it less hardware intensive. • Topological quantum computing : This approach uses topological qubits, which are based on the “To do a calculation with high accuracy, each one of the steps has to be extremely accurate, like one error in 10 billion,” Salfi said. Today’s qubits fall well short of that threshold, with many of the most advanced systems achieving an error rate of one error in a thousand computations. Baker described why this challenge is so persistent. “You can’t eliminate all possible sources of uncertainty on the quantum computer because quantum mechanics is inherently uncertain,” he said. As quantum systems store information in superpositions, small disturbances can corrupt results unless carefully managed. Classical computers tolerate noise through well established error correction. Quantum systems require far more stringent error correction that compensates for noise by encoding information redundantly. Baker emphasized that while this idea is conceptually familiar, its implementation is not. “It’s very similar to ideas of classical error correction,” he said. “You basically send the message several times and decode against noise.” The difference is that quantum information cannot simply be copied or observed without disturbance, meaning error correction multiplies hardware requirements. “Instead of using just one qubit to store information, you have several,” Baker explained. Creating a single reliable logical qubit requires hundreds of thousands of physical ones, compounding fabrication and control challenges. Electronics design presents another barrier. Many quantum systems operate at millikelvin temperatures, where conventional components behave differently. TYPES OF QUANTUM COMPUTERS

If we want to do a precise simulation of a molecule with about 120 atoms, you can do it on a classical computer. But it will need a hard drive about the size of the solar system, and it will have to run for about 1,000 times the age of the universe. Dr. Martin Laforest, Quantum physicist, partner at Quantacet

”

Controlling these systems requires rethinking assumptions common in classical engineering environments. The field is only beginning this theory-to-practice transition. “We’re at the beginning of the error-correction era,” Baker said. Industry roadmaps may project fault tolerant machines later this decade, but near-term systems must still operate with imperfections.

properties of anyons (a type of quasiparticle that exists only in two-dimensional systems), to store and process information. It is theoretically less sensitive to local errors due to their inherent resistance to environmental disturbances. This system is hypothetical as anyons have yet to be measured in a lab setting. • Measurement-based quantum computing : This method prepares a highly entangled “cluster state” and then computes via measurements. • Adiabatic quantum computing : This approach encodes a problem into a physical energy landscape and lets quantum physics find low energy solutions. Quantum annealing systems are a restricted version of adiabatic systems.

18

Summer 2026 Innovation