Quantum

Google's Willow chip shows quantum errors can shrink as the chip grows

A 105-qubit processor crossed a milestone researchers have chased for nearly 30 years: adding more qubits made the system more reliable, not less.

Share on X WhatsApp Facebook

In 30 seconds

  • Willow is Google Quantum AI's new 105-qubit superconducting chip.
  • Each time the team made its error-corrected 'logical qubit' bigger, the error rate roughly halved.
  • On a benchmark task it finished in under five minutes what Google estimates would take a top supercomputer 10 septillion years.

Quantum computers are fragile. Their basic units, qubits, lose their information when they are disturbed by heat, vibration or stray signals. For decades, adding more qubits also meant adding more errors, which made bigger machines less reliable.

Google’s Willow chip, announced in December 2024, changes that picture. The team grouped physical qubits into a single logical qubit using a method called the surface code. Then they grew that group step by step, from a 3×3 grid to 5×5 and then 7×7. Each step cut the logical error rate roughly in half.

Why “below threshold” is the headline

Physicists call this behaviour operating below threshold. Once a system is below threshold, you can, in principle, keep adding qubits to make errors as small as you like. It is the condition every large, useful quantum computer depends on. The result was published in the journal Nature.

The benchmark everyone quoted

Google also ran random circuit sampling, a test designed to be extremely hard for normal computers. Willow finished it in under five minutes. Google estimates the same job would take the Frontier supercomputer about 10 septillion (10²⁵) years.

That number is eye-catching, but the task itself is not useful for chemistry, medicine or finance. It is a stress test, not an application.

What comes next

Google’s roadmap needs logical qubits with far lower error rates and many more of them before machines can simulate molecules or break codes. Willow is the first clear sign that the engineering path to get there works.

Why it matters for India

India's National Quantum Mission aims to build machines with 50 to 1,000 physical qubits within eight years. Willow shows that the hard part is not counting qubits but correcting their errors, which is where Indian labs and startups will need to focus.

Our take

This is a real scientific milestone, but not a useful quantum computer yet. The benchmark Google used has no practical application. The result that matters is the error-correction curve, and that is the number to watch in every quantum announcement from now on.

Sources

We read these before writing this story. Go to them for the full detail.

  1. Google Quantum AI
  2. Google blog
  • Google
  • error correction
  • qubits
  • National Quantum Mission

Type to search every story. Use the arrow keys and Enter to open one.