Entanglement is a connection between two or more qubits where measuring one instantly determines the state of the other, no matter how far apart they are. It’s one of the core resources that gives quantum computers their power. Without it, a quantum computer couldn’t outperform a classical one on the problems that matter most.
To understand why entanglement matters for computing, it helps to first understand what it actually is, how engineers create it, and what makes it so difficult to maintain.
How Entanglement Works
A single qubit can exist in a superposition of 0 and 1 simultaneously. That’s useful on its own, but the real power emerges when two qubits become entangled. In an entangled pair, the qubits no longer have independent states. Instead, they share a single combined state, and the outcome you get from one qubit is always correlated with the outcome you get from the other.
Here’s a concrete example. Two entangled qubits can be set up so that when you measure them, you’ll either find both are 0 or both are 1, with a 50/50 chance of each. Before measurement, neither qubit is definitively 0 or 1. But the moment you measure one and find it’s a 0, the other is guaranteed to also be 0. This holds true whether the qubits are sitting next to each other on a chip or separated by over 1,200 kilometers, which is the current distance record set by China’s quantum satellite experiments.
This correlation isn’t like flipping two coins that were secretly pre-programmed. Physicists have tested this rigorously through experiments based on Bell’s inequality, a mathematical test that distinguishes quantum correlations from any possible pre-arranged (or “hidden variable”) explanation. The experiments consistently show that entangled particles violate Bell’s inequality, confirming that the correlations are genuinely quantum mechanical. Local realism, the intuitive idea that particles have definite properties before you look at them and can only be influenced by their immediate surroundings, is false.
How Entanglement Is Created
Engineers create entanglement using quantum logic gates, the building blocks of quantum circuits. The most common recipe uses two steps: first, put one qubit into superposition (an equal mix of 0 and 1), then apply a controlled-NOT gate (CNOT) between that qubit and a second one.
A CNOT gate works like a conditional switch. It has a control qubit and a target qubit. If the control qubit is 0, nothing happens to the target. If the control qubit is 1, the target flips. In classical computing, this is straightforward. But when the control qubit is in superposition, something new happens. The CNOT gate processes both possibilities at once, producing a final state where the two qubits are linked: the combination 00 and the combination 11 both exist simultaneously, but 01 and 10 do not. That output state cannot be separated back into two independent qubits. The qubits are now entangled.
Why Quantum Computers Need Entanglement
Superposition alone isn’t enough to make a quantum computer dramatically faster than a classical one. The key insight, proven mathematically, is that any quantum algorithm running on pure states needs entanglement across an increasing number of qubits to achieve an exponential speedup over classical computation. Without growing entanglement, a quantum system can be efficiently simulated on a regular computer, which defeats the purpose.
Entanglement is what lets qubits work together in a way classical bits cannot. In a classical computer, each bit is independent. Two bits store two pieces of information. But two entangled qubits share a joint state that encodes correlations no pair of classical bits can replicate. As you add more entangled qubits, the space of possible correlated states grows exponentially. This is the resource that algorithms like Shor’s factoring algorithm exploit. Researchers have explicitly identified the buildup of multi-qubit entanglement as Shor’s algorithm progresses through its steps.
Think of it this way: superposition lets each qubit explore multiple possibilities, but entanglement coordinates those explorations across qubits so they work in concert rather than independently. That coordination is what produces answers faster than any classical shortcut could.
Quantum Teleportation: Entanglement in Action
One of the clearest demonstrations of entanglement’s practical use is quantum teleportation, a protocol for transferring the exact state of a qubit from one location to another. It doesn’t move the physical qubit or transmit information faster than light, but it does accomplish something no classical method can: a perfect transfer of quantum information.
The protocol, developed at IBM and other research groups, works like this. Two people (traditionally called Alice and Bob) each hold one qubit from an entangled pair. Alice has a third qubit whose unknown state she wants to send to Bob. She performs a series of operations on her two qubits and then measures them, getting two ordinary bits of information (two 0s or 1s). She sends those two bits to Bob through a normal communication channel. Based on those two bits, Bob performs one of four simple operations on his qubit, and it ends up in exactly the state Alice’s original qubit was in.
The entangled pair gets “burned” in the process. After the protocol, Alice’s qubits are no longer entangled with Bob’s. The shared entanglement was consumed as a resource, much like fuel. This illustrates something important: entanglement isn’t free. It has to be created, maintained, and spent deliberately.
Why Entanglement Is So Fragile
The biggest engineering challenge in quantum computing is keeping qubits entangled long enough to finish a calculation. Entanglement is extremely sensitive to environmental interference, a process called decoherence. Any interaction with the surrounding environment (heat, vibration, stray electromagnetic fields) can cause qubits to lose their quantum states and the entanglement between them to collapse.
This is why quantum computers are typically cooled to temperatures colder than outer space and shielded from external noise. Even with those precautions, qubits lose coherence quickly, often within microseconds to milliseconds depending on the hardware. To combat this, researchers use quantum error correction codes, which spread information across extra qubits so that errors can be detected and fixed. Another technique, entanglement distillation, takes several weakly entangled pairs and combines them to produce fewer but more strongly entangled pairs. Both approaches add significant overhead in terms of the number of qubits required.
Current Scale of Entanglement
How many qubits can actually be entangled today? The answer depends on the type of system. In 2024, a photonic quantum computer named Aurora synthesized a cluster state (a specific entangled structure used for computation) spanning 86.4 billion modes across separate chips. That number sounds staggering, but photonic modes are different from the individually controllable qubits in superconducting or ion-trap systems, where entangling even a few hundred qubits with high fidelity remains a major achievement.
The gap between these numbers reflects a broader reality: generating entanglement at scale is one thing, but generating high-quality, controllable entanglement suitable for running algorithms is a harder problem entirely. Scaling up while keeping error rates low enough for practical computation is the central challenge the field is working to solve.