In the spring of 1935, in his office at the Institute for Advanced Study in Princeton, Albert Einstein was trying to break quantum mechanics. He had been trying for nearly a decade. He believed, with religious conviction, that the new physics being built by the younger generation in Copenhagen and Göttingen was incomplete. Beneath its probabilities, he thought, there had to be a deeper, deterministic reality, a hidden ledger on which the universe pre-recorded its outcomes the way relativity had pre-recorded the bending of starlight.

He had two postdoctoral assistants helping him. Boris Podolsky was a meticulous Russian-born theorist. Nathan Rosen was a young Brooklyn engineer turned physicist who had become Einstein's assistant at the Institute the year before. The three of them cooked up a thought experiment. Imagine, they wrote, two particles allowed to interact briefly and then sent in opposite directions. Quantum mechanics says that measuring the position of one forces the wavefunction of the other, however far away, to collapse into a definite position. Measuring the momentum of one forces the other to a definite momentum. The choice you make on Earth, in their analysis, instantly forces a state on a particle that may now be on Alpha Centauri.

This, the three argued, was absurd. Either there was a hidden communication channel propagating faster than light, which contradicted relativity, or the particles already carried "elements of reality" with them, which the quantum formalism failed to include. Either way, quantum mechanics was incomplete.

Podolsky drafted the paper. Einstein, reading the published version, was annoyed. He wrote to Erwin Schrödinger that summer that the essential point had been "smothered by formalism." But the title held: Can Quantum-Mechanical Description of Physical Reality Be Considered Complete? It appeared in Physical Review on 15 May 1935. It was probably the last great theoretical trap Einstein, Podolsky, and Rosen would set for Niels Bohr.

Bohr never admitted he had been caught. He published a counter-paper in the same journal that October, defending quantum mechanics as it stood. Most physicists shrugged, accepted the new theory worked, and got on with the practical business of nuclear reactors and transistors. The argument went dormant for a generation.

In March 1947 Einstein, now sixty-eight and still unconverted, wrote a private letter to his old friend Max Born. He could not seriously believe in quantum mechanics, he said, because the theory could not be reconciled with the idea that physics should represent a reality "free from spooky action at a distance." The original German was spukhafte Fernwirkung. He did not know the phrase would outlive him. He died in 1955 still believing he was right.

Nine years after Einstein's death, an unassuming Northern Irish particle physicist named John Stewart Bell, on a year's leave from his day job at CERN, picked the EPR argument back up. Bell had grown up in working-class Protestant Belfast. None of his three siblings had finished high school; his mother had pushed him to a technical secondary and then to Queen's University Belfast, where he took two consecutive first-class degrees, in experimental physics and then mathematical physics. His sabbatical took him through Stanford, Wisconsin, and Brandeis. Somewhere in there, he sat down and translated Einstein's philosophical complaint into a piece of mathematics.

If the universe was the way Einstein wanted, certain correlations between widely separated particle measurements would never exceed a particular bound. If quantum mechanics was right, those correlations would, occasionally, exceed it. Bell had turned a forty-year metaphysical dispute into something an experimentalist could measure with a polariser and a counter.

He sent the paper to a small new American journal called Physics Physique Физика, which printed in English, French, and Russian, paid its authors a small honorarium instead of charging them, and folded after only four years. Bell took the money and used it to buy reprints of his own paper, which he then posted to physicists around the world, hoping someone would notice. The paper appeared on 4 November 1964. For roughly a decade, almost nobody did.

The man who eventually noticed, properly, was a French graduate student named Alain Aspect, working at the Institut d'Optique in Orsay. By the late 1970s he had decided to do for Bell's inequality what his Parisian generation did best, which was to build a beautifully designed optical experiment around a single hard question. With Jean Dalibard and Gérard Roger he built a calcium-atom source that emitted pairs of entangled photons, sent them through fast-switching polarisers separated by metres of laboratory bench, and counted coincidences.

In December 1982 the team published their result in Physical Review Letters. Bell's inequality was violated, by five standard deviations. Einstein had been wrong about the kind of theory quantum mechanics was. There was no hidden mechanism the formalism failed to capture. There was, instead, a real and irreducible spookiness about the universe at the level of individual photons. Forty years later, in 2022, the Royal Swedish Academy of Sciences gave the Nobel Prize in Physics to Aspect, John Clauser, and Anton Zeilinger for that work and its descendants.

Zeilinger had, in 1997, with a small Innsbruck team, taken the next step. They borrowed a protocol Charles Bennett at IBM had proposed in 1993, and used a pair of entangled photons to transfer the polarisation state of a third "message" photon onto the partner of one of the entangled pair. They did not move any matter. They moved the state. In the language of the protocol, they teleported.

The Echo

This week, an international collaboration led by Klaus Jöns at Paderborn University and Rinaldo Trotta at Sapienza University in Rome announced, in Nature Communications, the result of about ten years of joint work. They had teleported the polarisation state of a single photon emitted by one solid-state quantum dot onto another, dissimilar quantum dot. Between the two dots was a hybrid network of optical fibre and a 270-metre open-air laser link strung between two buildings of the Sapienza campus. The fidelity of the transfer was 82 percent, more than ten standard deviations above the classical limit. The quantum dots had been grown in a clean room in Linz; their resonant cavities had been nano-fabricated in Würzburg.

The point of the result is not the distance. Two hundred and seventy metres is short by entanglement standards. Photon pairs have already been exchanged between Chinese satellites and ground stations more than twelve hundred kilometres apart. The point is that the two quantum sources were different. A real quantum internet has to teleport states between mismatched devices, the way the regular internet routes packets across mismatched routers in different cities. This is the year that began to be a serious engineering problem rather than a physics one.

What that engineering is for is not abstract. The quantum internet has three near-term jobs, and Sapienza's mismatched-source teleportation is the missing piece for all of them. The first is unconditionally secure communication: a cryptographic key distributed by quantum teleportation cannot be intercepted in flight without the act of interception leaving a measurable trace on the receiving end, which means that for the first time in human history banking, defence, and intergovernmental traffic could be carried on a channel that is eavesdrop-proof by physics rather than by mathematics. China demonstrated this at intercontinental scale in 2017, when its Micius satellite distributed shared keys between Beijing and Vienna and the two cities held a quantum-encrypted video conference. The second job is distributed quantum computing: a single chip cannot hold the millions of error-corrected qubits a useful quantum computer will need, so the only realistic path to scale is to link smaller processors together by teleporting quantum gates between them, which an Oxford-led group did between two trapped-ion modules in 2024. The third job is precision sensing: networks of atomic clocks and gravimeters linked by shared entanglement can resolve time and gravity to a finer scale than any single instrument, with applications from satellite navigation to dark-matter searches. Each of these waits on the same engineering primitive, the ability to move a quantum state cleanly from one device to a different device. Sapienza did that this week.

What Einstein hated about quantum mechanics, in 1935 and again in 1947, is not a philosophical embarrassment any longer. It is the protocol stack. The spooky action he refused to accept is the mechanism on which the next generation of secure communication is being built. The "elements of reality" he insisted must be there, the local hidden variables he was sure quantum theory was leaving out, are still missing. Aspect closed the door on them in 1982. Roman optical engineers are now exporting the consequences across a campus courtyard.

Einstein got the right answer to a slightly different question. Quantum mechanics is, in the way he meant it, incomplete. It does not hand you a tape on which the universe pre-records its measurements. But its non-locality is not a defect to be patched. It is a resource to be wired.

He spent his last twenty years convinced that nature could not be that strange. Nature, it turns out, is exactly that strange, and we have just begun to sell it in increments.