Entanglement in one coffee
On this page
Einstein disliked entanglement enough to call it "spooky action at a distance". Nearly a century later it's measured routinely in labs, it has won a Nobel Prize, and it sits at the heart of quantum computing. Here's what it is, what it isn't, and why it matters, in the time it takes to drink a coffee.
Two qubits that answer together
In Parts 1 and 2 we looked at single qubits, each holding its own blend of 0 and 1. You can also prepare two qubits so they share a single state instead of each having their own. That's entanglement.
The simplest version works like this. Measure either qubit and you get 0 or 1, completely at random, 50/50. But measure both and they always match. If the first says 0, the second says 0. If the first says 1, so does the second.
| Run | Qubit A | Qubit B | Match? |
|---|---|---|---|
| 1 | 0 | 0 | yes |
| 2 | 1 | 1 | yes |
| 3 | 1 | 1 | yes |
| 4 | 0 | 0 | yes |
| 5 | 1 | 1 | yes |
| 6 | 0 | 0 | yes |
Before measurement, quantum mechanics doesn't assign either qubit a value of its own; it describes only the pair. The pair holds the blend "both 0, or both 1", with an amplitude of about 0.71 on each option. There's no amplitude at all on "0 and 1" or "1 and 0", so those mixed results never appear. Measure one, and you can predict the other's result with certainty, however far apart the two qubits are.
This is something a pair of separate qubits simply can't do. Two independent qubits each in a 50/50 blend would give you matching results only half the time. Entanglement means the pair has to be described as one thing.
One caution before going further. This perfect matching, on its own, doesn't prove anything strange. As the next sections show, pre-arranged instructions could produce it too. The evidence that entanglement is genuinely different comes from measuring the pair in several different ways and comparing the results. That's what Bell's test does.
How do you entangle two qubits?
It sounds exotic, but inside a quantum computer it takes just two steps, both standard operations.
Step one. Start with two qubits, both plain 0. Put the first one into an even blend of 0 and 1, exactly as in Part 2. The second qubit is still a plain 0.
Step two. Apply a "controlled flip": an operation that flips the second qubit from 0 to 1, but only if the first qubit is 1. If the first qubit were an ordinary bit, this would be simple: either it's 1 and the second flips, or it's 0 and nothing happens.
But the first qubit is a blend. So the operation does both, each with its own amplitude. Part of the state becomes "first 0, second 0" and part becomes "first 1, second 1". The two qubits are now tied together. Neither has a value of its own, and their results will always match. That's the entangled pair from the table above.
Every quantum computer can do this, and the controlled flip is one of the most-used operations in quantum programs. It's also one of the most error-prone, because it requires two delicate qubits to interact precisely while staying isolated from everything else. When companies quote error rates, the two-qubit error rate is usually the number that matters most.
Entangling particles far apart, as in long-distance experiments, works on the same idea. The two particles are entangled close together first, then carefully separated, usually as photons sent down optical fibres or through the air.
Isn't that just a pair of gloves?
It's the obvious objection. Put a left glove and a right glove in two boxes, mail them to opposite ends of the country, open one, and you instantly know what's in the other. Nothing spooky there. The answer was decided when the boxes were packed.
Einstein made essentially this argument. In 1935, with Boris Podolsky and Nathan Rosen, he argued that quantum theory must be incomplete: the particles must carry hidden instructions, like gloves, that we just haven't found yet. For almost thirty years nobody could see a way to test it.
Bell's test, as a game
In 1964 the physicist John Bell found a way. The easiest version to understand is a game for two players, call them Asha and Ben.
Asha and Ben sit in separate rooms and can't communicate. Each round, a referee hands each of them a card with a random letter, either X or Y. Each must answer with 0 or 1. They win the round if their answers match, except when both were handed Y. In that case they win only if their answers are different.
They can agree a strategy beforehand, which is the gloves idea: pre-packed instructions. But whatever plan they agree, they can't win more than 75% of rounds on average. Try it: "always answer 0" wins every round except the Y-and-Y ones, which turn up a quarter of the time. No cleverer plan does better. Three out of four is the ceiling for any pre-arranged strategy.
Now give Asha and Ben one entangled qubit each. Depending on their letter, each measures their qubit in a particular way and reports the result. Quantum mechanics predicts they win about 85% of rounds. That's above the 75% ceiling, and no hidden instructions in any gloves could manage it.
Real experiments run this game, or close cousins of it, with photons or atoms. John Clauser ran an early version in 1972. Alain Aspect's experiments in the early 1980s tightened it, and in 2015 several labs closed the remaining loopholes. Results kept beating the limit. Aspect, Clauser and Anton Zeilinger shared the 2022 Nobel Prize in Physics for this work. Entangled qubits really are not gloves.
What entanglement can't do
It can't send messages faster than light. This is the most common misunderstanding, so it's worth being precise.
If you hold qubit A, all you ever see is random 0s and 1s. Nothing your friend does to qubit B, no matter how far away, changes what that randomness looks like to you. The agreement only becomes visible when the two of you compare notes, and comparing notes needs an ordinary phone call, email or courier. No useful information travels through entanglement on its own.
Entanglement is a shared correlation, not a communication channel.
The same rule limits "quantum teleportation", another headline favourite. Teleportation moves the state of a qubit from one place to another using an entangled pair, without sending the qubit itself. But it needs two ordinary bits of information sent the normal way, and it destroys the original state in the process. Nothing travels faster than light, and no people are being beamed anywhere.
Why it matters in practice
Computing. Without entanglement, a set of qubits is just a row of independent coins, and an ordinary computer could simulate them easily. Entanglement lets amplitudes combine across the whole group, which is what makes the doubling from Part 2 useful. Most quantum algorithms create and use large amounts of entanglement along the way.
Secure communication. Some quantum key distribution schemes use entangled pairs to create shared secret keys. Because measuring disturbs a quantum state, an eavesdropper leaves detectable traces. In 2017 a Chinese satellite called Micius sent entangled photons to ground stations about 1,200 km apart. This is a different idea from post-quantum cryptography, the main defence banks are adopting, which we'll meet in Part 5.
Sensing. Entangled states can make some measurements of time, magnetic fields and gravity more precise than ordinary methods allow. Atomic clocks and some medical and navigation sensors are early beneficiaries.
Common questions
Does measuring one qubit "send a signal" to the other? No signal is sent that anyone could use. The results are correlated, but each side alone sees pure randomness.
How long does entanglement last? As long as the qubits stay isolated. Any stray interaction with the outside world tends to spread the entanglement into the environment and destroy it, which is one more reason qubits are so fragile.
Can more than two qubits be entangled? Yes. Any number of qubits can share one entangled state, and useful quantum computations entangle many qubits at once. Researchers have created entangled states across dozens of qubits on today's machines. The more qubits involved, the harder the state is to keep clean, because there are more ways for the outside world to interfere.
Is entanglement rare in nature? No. Particles become entangled whenever they interact. What's rare and difficult is keeping entanglement clean and controlled long enough to use it.
- Entangled qubits share one state: each result is random, but the results are linked.
- Bell-style experiments show the link is stronger than any pre-arranged plan could explain.
- It can't carry messages on its own, so it doesn't break the speed of light.
Sources and further reading
Tags show what kind of source each one is. A standard or government guidance is an official document; a peer-reviewed paper has been checked by other experts; a preprint has not been peer-reviewed yet; an experiment reports a real-world demonstration; a company announcement is the company's own account. Dates and figures were checked against these sources on 11 October 2026. Spotted an error? Email hello@plainquantum.com and it will be corrected, with a note.
- Peer-reviewed paperCan quantum-mechanical description of physical reality be considered complete?Einstein, Podolsky and Rosen, Physical Review, 1935
- Peer-reviewed paperOn the Einstein Podolsky Rosen paradoxJohn Bell, Physics, 1964
- Peer-reviewed paperProposed experiment to test local hidden-variable theories (the CHSH game)Clauser, Horne, Shimony and Holt, Physical Review Letters, 1969
- ExperimentLoophole-free Bell inequality violation using electron spins separated by 1.3 kilometresHensen et al., Nature, 2015
- ReferenceThe Nobel Prize in Physics 2022Nobel Prize Outreach, 2022
- ExperimentSatellite-based entanglement distribution over 1200 kilometersYin et al., Science, 2017
- Peer-reviewed paperTeleporting an unknown quantum state via dual classical and Einstein-Podolsky-Rosen channelsBennett et al., Physical Review Letters, 1993
- BookQuantum Computation and Quantum Information (10th anniversary edition)Michael Nielsen and Isaac Chuang, Cambridge University Press, 2010