Google’s “Willow” Quantum Processor & Verified Quantum Advantage

In October 2025, researchers from Google, MIT, Stanford, and Caltech published two papers in Nature demonstrating a verifiable quantum advantage using Google’s Willow Quantum Processor.

This marks a major step in quantum computing showing performance beyond classical supercomputers for specific problems.

Based on quantum interference like waves meeting:

• Constructive interference: wave crests reinforce each other → amplifies correct answers.

• Destructive interference: crests and troughs cancel → removes wrong answers.

Quantum computers use qubits (quantum bits) that exist in multiple states simultaneously (superposition).

The goal is to engineer interference so that measurement yields the correct (optimal) answer with high probability.

Introduced a quantum algorithm for solving optimisation problems.

Uses quantum Fourier transform to manipulate wave-like qubits.

Good solutions → constructive interference;
Bad solutions → destructive interference.

Demonstrated for the “Optimal Polynomial Intersection Problem”, where DQI outperformed any known classical algorithm in finding approximate solutions faster.

Scrambling = process by which information spreads across a quantum system due to entanglement and interactions.

Analogy: dropping blue dye in a pool → initially local, then spreads until uniformly mixed.
Information isn’t lost just hidden in correlations between particles.

Measured scrambling using OTOC (Out-of-Time-Order Correlator) — a way to quantify how information disperses.

The “warehouse” analogy (sound echo + bell) demonstrates how scientists detect faint imprints of scrambled data through interference measurements.

Willow Quantum Processor solved circuits that would take the world’s 2nd fastest supercomputer >3 years Willow did it in ~2 hours.

Demonstration was verifiable, unlike Google’s earlier 2019 claim (random circuit sampling).

Google claimed “quantum supremacy” in 2019 using its Sycamore processor for random circuit sampling.

However, results weren’t scientifically verifiable — only statistically consistent.

2025’s Willow demonstration is verifiable since results can be checked by classical simulations or other quantum computers.

First verifiable quantum advantage claim for a scientifically meaningful problem.

Opens paths for:

• Hamiltonian learning (inferring physical system parameters)

• Optimization problems

• Quantum simulation

Still in early stage — practical applications depend on error correction, scaling, and stability of qubits.

Nobel Laureate Michel Devoret, Chief Scientist at Google Quantum AI, contributed foundational work on superconducting quantum circuits, key to processors like Willow.