top of page
Domains
Quantum Experiment Produces Certifiably Perfect Randomness
A quantum experiment transforms imperfect randomness into mathematically certified random numbers, demonstrating a uniquely quantum capability with major cybersecurity applications.
Kulikov et., al
May 27, 2026

In May 2026, researchers at ETH Zurich demonstrated a fundamentally quantum capability with direct implications for cybersecurity: transforming imperfect randomness into certifiably ideal random numbers.
The work, published in Nature on May 27, experimentally implemented a process known as randomness amplification using entangled superconducting qubits and a loophole-free Bell test.
Random numbers are a critical foundation of modern cryptography. Encryption keys, digital authentication and secure communications all depend on generating values that attackers cannot predict. However, practical random-number generators inevitably contain small biases or imperfections.
Randomness amplification starts with a source that is only partly random and uses quantum correlations to produce a new sequence whose randomness can be mathematically certified.
The ETH team achieved this using two superconducting quantum circuits and a high-performance Bell-test experiment. The Bell inequality violation provides evidence that the correlations being measured cannot be reproduced by a classical local model. The resulting random bits can therefore be certified without requiring complete trust in the internal operation of the devices—a concept known as device independence.
The theoretical importance is particularly striking because randomness amplification has been proven impossible using purely classical information processing. Nature therefore describes the experiment as a definitive form of quantum advantage: the quantum system accomplishes an information-processing task that cannot be achieved classically under the same assumptions.
Potential applications include cryptographic key generation, secure digital identities, public randomness services, lotteries and blockchain systems. ETH researchers have compared the concept to an atomic clock for randomness—a trusted physical standard that other systems could rely upon.
This differs significantly from headline-grabbing quantum-computing speed claims. The experiment does not solve a massive optimization or chemistry problem.
Instead, it demonstrates a smaller but conceptually unambiguous application in which quantum physics provides a capability that classical systems fundamentally cannot duplicate.

bottom of page