A team led by Renato Renner of the Quantum Information Theory Group and Andreas Wallraff of the Experimental Physics Group at ETH Zurich has published a paper in the journal Nature, reporting the first physical experiment to generate and rigorously certify “perfect randomness” — a sequence of bits that can be proven to remain random and unpredictable under any method of analysis. The experimental setup consists of two superconducting quantum chips connected by a 30-meter-long cryogenic tube cooled to near absolute zero. The two chips form quantum entanglement by exchanging microwave photons back and forth; the 30-meter physical separation ensures that during the measurement process, even information traveling at the speed of light cannot be exchanged between the two ends, thereby meeting the strict “locality” requirements of a Bell inequality test. The research team fed the initial randomness produced by a flawed random number generator into the Bell test, then applied a specially designed algorithm to perform “randomness amplification” on the measurement results, ultimately outputting a perfectly random sequence that can be mathematically certified. Renner likened this breakthrough to “crossing a ridge” — technical improvements have allowed the team, for the first time, to reach a critical threshold that existed theoretically but had remained experimentally inaccessible.
Previously, random number generators based on quantum physical phenomena such as shot noise and avalanche diodes, while widely used in cryptography, could not mathematically rule out tiny systematic biases or potential backdoors, achieving only “practical randomness” rather than “certified perfect randomness.” The research team stated that the long-term significance of this result is that it could provide a physical infrastructure for digital security akin to the role of atomic clocks for timekeeping — a physically certified source of randomness that other systems can trust, with potential applications covering scenarios such as sensitive communication encryption, digital identity, and public randomness services (e.g., lotteries and blockchain applications).