Google’s Willow chip shows quantum error correction getting better as it scales
In short: on 9 December 2024 Google announced Willow, a 105-qubit superconducting chip. Its main result, published in Nature, is that errors went down as Google made its error-corrected qubits larger. This “below threshold” behaviour is a key requirement for building large, reliable quantum computers.
What happened
Google Quantum AI tested grids of physical qubits of increasing size (3×3, 5×5 and 7×7) working together as a single error-corrected logical qubit. Each step up in size roughly halved the error rate. Until now, adding more qubits usually added more errors, so this was the first convincing demonstration that error correction can improve as a system scales.
Google also reported that Willow completed a “random circuit sampling” benchmark in under five minutes, a task it estimates would take one of today’s fastest supercomputers about 1025 years. The benchmark has no practical use; it is a stress test that classical computers find very hard.
Why it matters
- Error correction is the bottleneck. Breaking RSA or simulating complex molecules needs logical qubits that stay reliable for billions of operations. Willow shows the approach can work at small scale.
- It does not threaten encryption today. Willow has 105 physical qubits; current estimates for breaking RSA-2048 need close to a million.
- It strengthens the case for planning now. Milestones like this are why regulators set migration deadlines around 2030 and 2035.
What it means for Hong Kong and New Zealand
Nothing changes immediately, but boards that ask “is quantum computing real?” now have a clear, peer-reviewed answer. It is a good moment to start, or refresh, a cryptographic inventory.
- Google, Meet Willow, our state-of-the-art quantum chip (9 December 2024)
- Google Quantum AI, Quantum error correction below the surface code threshold, Nature (December 2024)
