Quantum Entanglement
Quantum entanglement is a correlation between two or more qubits that cannot be explained by either qubit having a definite state of its own — measure one qubit of an entangled pair and the outcome instantly constrains what a measurement on the other qubit will show, no matter how physically far apart the two devices sit. A classical correlation works differently: two sealed envelopes, one with a red card and one with a blue card, are "correlated" the moment they're sealed — opening one just reveals a fact that was already fixed. An entangled pair of qubits has no such pre-assigned answer; the joint state |00⟩ + |11⟩ (normalized) genuinely does not specify what either qubit's individual measurement will return until one of them is measured, at which point both snap to a definite, correlated outcome. Einstein called this "spooky action at a distance" and tried to argue it meant quantum mechanics was incomplete — Bell test experiments conducted since the 1970s, and especially the 2022 Nobel-Prize-winning work of Aspect, Clauser, and Zeilinger, confirmed the correlations are real and cannot be reproduced by any theory with pre-assigned local values.
On real superconducting hardware, entanglement isn't a separate exotic step — it's the direct output of a two-qubit gate applied to two coupled transmons. IBM's processors and Google's Sycamore-derived chips both lean on the cross-resonance gate: drive one qubit at the other qubit's resonant frequency through their shared coupling (a capacitor or a tunable coupler bus), and the interaction Hamiltonian entangles the pair as a direct consequence of the physics — no separate "entangling mechanism" bolted on afterward. Rigetti and other labs instead favor the iSWAP gate family, tuning two transmons' frequencies into resonance with a flux-tunable coupler so that a single excitation swaps coherently between them, which is itself an entangling operation when stopped halfway. IonQ's trapped-ion qubits entangle through a different physical channel entirely — a laser-driven Mølmer-Sørensen gate couples two ions' internal states through their shared vibrational motion in the trap — proof that entanglement is a statement about the joint quantum state, not about any one piece of hardware that produces it.
Entanglement is also the resource that separates a genuine quantum algorithm from a device that just runs lots of independent coin flips in parallel. Shor's factoring algorithm and Grover's search both rely on entangling the qubits that hold the problem's input with the qubits that accumulate the answer, so that interference effects can act on the whole joint state rather than on isolated, unconnected qubits — without entanglement, an n-qubit register behaves mathematically like n separate classical bits and no exponential speedup is possible. This is also why entanglement is the quantity benchmarked most closely on new hardware: Google Quantum AI's 2019 "quantum supremacy" claim and IBM's annual roadmap updates both report two-qubit gate fidelity (currently above 99.5% on the best superconducting pairs) because every entangling operation introduces a small chance of error, and algorithms that need deep circuits of many chained entangling gates die quickly if that fidelity isn't high enough.
Read an entangled pair not as "two qubits that know about each other" in some mysterious sense, but as a single joint quantum object described by one wavefunction that happens to span two physically separate pieces of hardware — the entangling gate is the one operation that actually writes correlations into that object, and a quantum processor's two-qubit gate fidelity is the real ceiling on how much useful entanglement it can generate before noise unravels it.