Google has unveiled its new Willow quantum processor, claiming a major leap in quantum error correction—a key step toward building large-scale quantum computers capable of solving practical problems. Quantum processors today are highly prone to errors, and that vulnerability magnifies as they grow larger.
Error-correction techniques combine many “physical qubits” into one “logical qubit,” making the system more robust against mistakes. But this strategy only works if each qubit’s baseline error rate remains below a critical threshold.
In Nature, Google details how Willow’s qubits, arranged in ever-larger arrays, cut the overall error rate by half each time they increased the grid size. These logical qubits lasted more than twice as long as the physical ones, demonstrating that Google’s hardware improvements—better fabrication methods and circuit optimizations, significantly boosted qubit lifetimes from roughly 20 to 68 microseconds.
Besides highlighting Willow’s error-correction prowess, Google ran a contrived test that took under five minutes on the chip but would supposedly take the world’s second-fastest supercomputer, Frontier, an astronomical amount of time. Although this test doesn’t solve a useful problem, it reveals Willow’s speed potential.
Despite the encouraging progress, there’s still a daunting road ahead. Most real-world challenges will require logical qubits with much lower error rates, meaning each logical qubit must be made of around 1,000 physical qubits. Google’s demonstration also only stored information; executing complex calculations at scale would require billions of logical operations. Still, Google’s results suggest that fully functional, large-scale quantum computers may be closer to reality than before.