Glossary of Essential Quantum Computing Terms for Beginners

Quantum computing comes with a dense vocabulary of its own. This glossary explains the terms you will run into most often, from qubits and superposition through gate fidelity, error correction, and the different kinds of quantum hardware, in plain language. Jump to a letter below, or scroll through the whole thing.

#

2-Qubit Gate Equivalent (2QGE): A normalized way to measure how much computational work a quantum processor has done, expressed as an equivalent number of two-qubit gates. Because gate duration and quality differ across hardware platforms, 2QGE lets you compare workloads on a common scale. A native gate on a given machine counts as a fraction or multiple of a 2QGE depending on how its duration compares to that machine's standard two-qubit gate. It is mainly used for capacity planning and cross-platform comparison rather than describing a single circuit.

A

Algorithmic Benchmarking: A way of testing quantum hardware using algorithms or subroutines that resemble real applications, rather than purely random circuits. The goal is to measure performance on the kind of problem structure a device would actually be asked to solve, similar to how LINPACK benchmarks classical supercomputers with real linear algebra. Results are usually reported as success probability or solution quality on a specific task. This tends to correlate more closely with practical usefulness than benchmarks built from random circuits alone.
Algorithmic Qubit: A metric meant to capture how many qubits in a system are actually usable for a computation, rather than simply counting physical qubits. Noise, limited connectivity, and gate errors reduce the number of operations that can be reliably chained together, so not every physical qubit contributes equally to a working algorithm. The count reflects roughly the largest circuit a device can run with results you can still trust. It is one of several industry attempts to describe hardware capability more honestly than raw qubit count.

B

Barium Qubit: A qubit built from a single trapped barium ion rather than the more common ytterbium ion. Barium has a more complex internal energy structure, which can support higher native gate and readout fidelities when the ion is controlled precisely. It also responds mainly to visible-wavelength light, opening the door to more standard fiber-optic components in the control system. Some ion-trap teams are exploring barium as a way to push fidelity and manufacturability further than earlier ion species allowed.
Below-Threshold Error Correction: The regime in which adding more physical qubits to an error correction code makes the encoded logical qubit more reliable rather than less. Every code has a threshold physical error rate; only when the hardware operates below it does increasing the code distance drive the logical error rate down, ideally exponentially. Reaching this regime is the practical sign that quantum error correction is working as intended rather than adding overhead for no net benefit. It was first demonstrated convincingly on a superconducting processor in 2024, and is widely regarded as a turning point on the path to fault-tolerant machines.
Bloch Sphere: A geometric picture used to visualize the state of a single qubit as a point on the surface of a sphere. The north pole represents |0>, the south pole represents |1>, and every other point on the surface represents a valid superposition of the two. Single-qubit gates correspond to rotations of this point around the sphere, which makes the Bloch sphere a useful tool for building intuition. It only works for one qubit in a pure state; entangled multi-qubit states cannot be drawn this way.

C

Circuit Depth: A measure of how many entangling gates occur in sequence within a quantum circuit, independent of how many qubits the circuit touches. A circuit with thirty two-qubit gates chained one after another has a depth of thirty, whether it acts on three qubits or three hundred. Depth matters because every additional gate gives noise another opportunity to accumulate before the final measurement. Comparing depth across circuits gives a rough sense of how demanding a computation is on a given piece of hardware.
Circuit Width: The number of qubits a circuit entangles together, independent of how many gates it uses to do so. A circuit is said to have a width of six if it entangles six qubits, regardless of whether it applies five gates or five hundred. Width and depth together describe the shape of a circuit, and both interact differently with a real device's noise. A wide, shallow circuit and a narrow, deep circuit can be equally hard to run well, just for different physical reasons.
Classical Computer: Any computer that stores and processes information using classical physics, where each bit is definitely a 0 or a 1 at all times. Most classical computers implement this with transistors, representing a bit as the presence or absence of electrical current in a semiconductor switch. Every operation is ultimately built from simple logical combinations of these binary states. Quantum computers are described in contrast to classical ones because they exploit superposition and entanglement, properties classical bits do not have.
CNOT Gate: A two-qubit logic gate that flips the state of one qubit, the target, only when a second qubit, the control, is in the state |1>. It is the standard building block for creating entanglement and, combined with a small set of single-qubit rotations, forms a universal gate set capable of expressing any quantum computation. Because it links two qubits directly, the CNOT gate is usually the noisiest and slowest operation on real hardware, which is why two-qubit fidelity is watched so closely.
Coherence Time: How long a qubit can hold onto its quantum properties, such as superposition, before its state is disturbed by the environment. It is often split into two numbers: T1, how long a qubit retains its energy state, and T2, how long it retains phase information. Longer coherence times give an algorithm more room to complete its gates before the qubit's information degrades. Trapped-ion qubits are notable for coherence times measured in seconds or minutes, far longer than many solid-state alternatives.
Computational Complexity: A way of describing how the resources a problem needs, typically time or memory, grow as the problem gets larger. Some problems scale gently and stay solvable at large sizes, while others scale exponentially and become impossible for any classical computer past a certain size. A subset of these hard-scaling problems can be solved efficiently on a quantum computer instead, which is where quantum computing's potential advantage comes from. A problem's complexity class is usually the first thing to check before assuming quantum hardware would help.
Connectivity: A description of which qubits in a device can interact directly through a two-qubit gate. Trapped-ion systems typically offer all-to-all connectivity, meaning any qubit can be entangled directly with any other qubit in the same trap. Many other platforms only allow gates between physically neighboring qubits, so entangling two distant qubits requires a chain of swap operations that adds noise and time. Higher connectivity generally means a compiler has fewer compromises to make when mapping an algorithm onto real hardware.
Crypto-Agility: The property of a system that lets its cryptographic algorithms be replaced without re-architecting the system around them. It matters most in the transition to post-quantum cryptography, where organizations must be able to swap classical key exchange and signature schemes for quantum-resistant ones, and to do so more than once as standards evolve. The goal is to treat cryptographic migration as a repeatable operation rather than a single disruptive event, which is why government guidance and standards bodies now name crypto-agility as an explicit requirement rather than an afterthought.

D

Decoherence: The gradual loss of a qubit's quantum information caused by unwanted interaction with its surrounding environment. It is the central obstacle between today's noisy hardware and a fully reliable quantum computer, since it erodes the superposition and entanglement that quantum algorithms depend on. Engineers fight decoherence with better isolation, colder operating temperatures, and increasingly with error correction codes that spread information across multiple physical qubits. The rate of decoherence relative to gate speed is one of the most important numbers describing any qubit technology.
Dilution Refrigerator: A specialized cooling apparatus that brings superconducting quantum processors down to around 10 to 20 millikelvin, colder than deep space. It reaches these temperatures by exploiting the mixing behavior of helium-3 and helium-4 isotopes across several cooling stages. The gold, chandelier-like structures seen in many quantum computing photos are the internal wiring and cooling stages of one of these refrigerators. Superconducting qubits require this extreme cold to hold their quantum properties, which is one reason that platform is harder to scale outside a single lab setup.
DiVincenzo Criteria: A set of five conditions, proposed by physicist David DiVincenzo in 2000, that a physical system must satisfy to be considered a viable quantum computer. In short, it must be scalable to many qubits, allow qubits to be initialized to a known state, support a universal set of quantum gates, allow individual qubit measurement, and keep coherence times long enough to complete a computation. These criteria are still used today as a checklist when evaluating new qubit technologies, from trapped ions to superconducting circuits to photonics.

E

Entanglement: A quantum mechanical correlation between two or more particles such that their combined state cannot be described by treating each particle separately, even when they are physically far apart. Measuring one entangled particle instantly tells you something about the state of its partner, a connection that has been experimentally verified many times. Entanglement is a resource that quantum algorithms actively create and consume; without it, a quantum computer behaves essentially like a fancy classical bit register. It also underlies related technologies such as teleportation and quantum key distribution.
Error-Corrected Qubit: Another name for a logical qubit: a group of physical qubits combined using an error correction code so the group behaves as one much more reliable qubit. Individual physical qubits are noisy and error-prone, but spreading information redundantly across many of them lets errors be detected and corrected without ever directly measuring, and destroying, the encoded quantum state. The number of physical qubits needed per error-corrected qubit depends on the code and target error rate, but commonly runs into the hundreds or thousands.
Evaporated Glass Trap (EGT): A proprietary ion-trap fabrication approach that enables multiple compute zones on a single chip. By allowing tighter confinement of the trapped ions and reducing unwanted heating, the design supports more precise control over each ion and makes it more practical to operate several ion chains in parallel. It is one example of how ion-trap hardware has moved from single research setups toward chip-scale, multi-core architectures, showing that hardware progress often comes as much from clever trap engineering as from the underlying physics.

F

Fault-Tolerant Quantum Computing: The ability of a quantum computer to keep producing correct results even though its individual physical components are imperfect and noisy. It requires enough qubits, sufficiently high-quality qubits, and an error correction scheme that stops small errors from snowballing into a ruined computation, all working together. Reaching fault tolerance is widely seen as the milestone that unlocks quantum computing's largest promised applications, since only then can circuits run long and deep enough to solve genuinely hard problems. The first building blocks arrived in 2024 to 2026, with below-threshold error correction demonstrated and commercial systems running tens of logical qubits, so today's devices are at the start of the fault-tolerant transition: useful for research and early error-corrected experiments, not yet reliable for long computations.

G

Gate Fidelity: A measure of how closely a real quantum gate matches its ideal mathematical operation, usually expressed as a percentage. A fidelity of 99 percent is the same as saying the gate has a 1 percent error rate on average. Two-qubit gates are typically far noisier than single-qubit gates and are usually the limiting factor for how deep a circuit can run before errors dominate the result. When comparing fidelity numbers between vendors, check whether the figure is a best case, an average, or something else entirely.
Gate Speed: How quickly a quantum processor can execute a single logic gate, typically measured in microseconds or nanoseconds depending on the platform. In today's noisy devices, the practical requirement is that every gate in an algorithm finish well within the qubits' coherence time, or the information decays before the computation completes. Different qubit technologies trade gate speed against other qualities; trapped ions, for example, tend to run slower gates than superconducting qubits but hold their state far longer, which often evens things out in practice.
Grover's Algorithm: A quantum search algorithm that finds a target entry in an unsorted list of N items using roughly the square root of N steps, compared to the N steps a classical computer needs on average. It works by repeatedly applying an operation that amplifies the probability of measuring the correct answer while suppressing the rest. The speedup is smaller than Shor's algorithm offers for factoring, but Grover's algorithm applies to a much broader range of problems, including search, optimization, and some machine learning tasks.

H

Hadamard Gate: One of the most commonly used single-qubit gates, taking a qubit starting in |0> into an equal superposition of |0> and |1>. Applied to a qubit already in superposition, it can also reverse the operation and return the qubit to a definite state. Because it is the simplest way to create superposition, the Hadamard gate appears near the start of a huge number of quantum circuits, setting up the parallel exploration of possibilities that gives quantum algorithms their potential advantage.
Harvest Now, Decrypt Later: A threat model in which an adversary records encrypted data today and stores it, intending to decrypt it later once a quantum computer capable of breaking today's public-key cryptography becomes available. Also called store-now-decrypt-later, it means any information that must stay confidential for many years is effectively exposed in the present, even though no machine can break the encryption yet. It is the main reason organizations are advised to migrate to post-quantum cryptography well ahead of the threat, and the relevant comparison, sometimes called Mosca's inequality, weighs how long data must stay secret plus how long migration takes against how soon a capable machine might arrive.
Hybrid Quantum-Classical Computing: An approach where a quantum processor and a classical computer work together on the same problem, each handling the part it is best suited for. Typically the quantum computer prepares and measures a parameterized circuit, while a classical optimizer analyzes the results and adjusts the circuit's parameters for the next round. This loop repeats until the result converges on a good answer. Most near-term quantum algorithms, including VQE and QAOA, are built this way because today's hardware is not yet reliable enough to run a full algorithm on its own.

I

Ion Trap: The core piece of hardware in a trapped-ion quantum computer: microfabricated electrodes arranged to generate an electromagnetic field that holds a line of charged atoms, ions, in place. A useful comparison is a maglev train, where the trap plays the role of the track that levitates and positions the train without touching it. Lasers or microwaves are then aimed at the trapped ions to perform gates and read out results. Modern ion traps are built using fabrication techniques adapted from the semiconductor industry.

L

Logical Qubit: A single, more reliable unit of quantum information built by combining many physical qubits under a quantum error correction code. Because individual physical qubits are error-prone, a logical qubit spreads information redundantly across a group of them so errors can be caught and corrected before they ruin a computation. The number of physical qubits needed per logical qubit varies by code: surface-code estimates once ran from the hundreds into the low thousands, but low-density parity-check codes and neutral-atom architectures have pushed the ratio far lower, and by 2025 to 2026 commercial devices were demonstrating tens of logical qubits at overheads close to two physical qubits per logical qubit. Counting logical qubits, not physical ones, is the more honest measure of real computing power.

M

Multi-Core QPU: A single quantum processor built with multiple separate compute zones on one chip, similar in spirit to a multi-core classical processor. Each zone can perform its own operations in parallel, and ions or other qubit carriers can be physically moved between zones to create entanglement across the whole device. This design is one route hardware makers are pursuing to scale up qubit counts without needing every qubit to interact directly with every other qubit at all times.

N

Neutral Atom Quantum Computing: An approach that holds individual, electrically neutral atoms in place using tightly focused laser beams known as optical tweezers, arranged in flexible patterns. Qubits are encoded in the atoms' internal energy states, and entangling gates are performed by briefly exciting pairs of atoms to high-energy Rydberg states where they strongly interact. Because tweezer arrays can be reconfigured and scaled relatively easily, this platform has produced some of the largest qubit counts demonstrated to date, while running the atom array itself at room temperature.
NISQ Era: Short for Noisy Intermediate-Scale Quantum, a term describing the generation of quantum hardware that dominated the late 2010s and early 2020s. NISQ devices have tens to a few hundred qubits, are not protected by full error correction, and accumulate enough noise that only relatively shallow circuits run reliably. Despite these limits, NISQ machines proved useful for developing and testing algorithms, running hybrid quantum-classical workloads, and probing early signs of quantum advantage on narrow problems. The field has now begun moving beyond the NISQ era, with the first below-threshold error correction and small error-corrected, logical-qubit systems demonstrated in 2024 to 2026.
NIST Post-Quantum Standards: The first set of post-quantum cryptography standards finalized by the United States National Institute of Standards and Technology in 2024, after a multi-year public competition. The core standards are ML-KEM (FIPS 203, for establishing shared keys), ML-DSA (FIPS 204, for digital signatures), and SLH-DSA (FIPS 205, a hash-based signature scheme kept as a hedge on different mathematical assumptions), with a further signature standard based on the FALCON scheme in preparation. Most rely on the hardness of structured lattice problems, which are not currently known to be vulnerable to quantum attack, and they are now being integrated into protocols, browsers, and hardware, commonly in hybrid combinations with the classical schemes they are set to replace.

P

Photonic Quantum Computing: An approach that uses individual photons, particles of light, as qubits, encoding information in properties like polarization, path, or arrival time. Because photons need no extreme cooling and travel naturally through optical fiber, this platform has a built-in advantage for quantum networking and room-temperature operation. The tradeoff is that generating, routing, and detecting single photons reliably enough for large-scale computation is a significant engineering challenge in its own right, which is why several companies pursue this route specifically for its fit with existing fiber infrastructure.
Physical Qubit: The actual physical hardware element used to store one unit of quantum information, whether a trapped ion, a superconducting circuit loop, a single photon, or another system entirely. Physical qubits are inherently noisy and subject to decoherence, gate errors, and imperfect measurement. When a company advertises a qubit count for its processor, that number almost always refers to physical qubits, a very different and much larger figure than the number of logical, error-corrected qubits those physical qubits could ultimately support.
Post-Quantum Cryptography: A family of cryptographic algorithms designed to remain secure even against an attacker with access to a large, fault-tolerant quantum computer. It exists because Shor's algorithm could, in principle, break the assumptions behind widely used systems like RSA once sufficiently powerful quantum hardware exists. Standards bodies have been evaluating and formalizing new algorithms based on different mathematical problems that are believed to resist both classical and quantum attacks. Organizations are encouraged to migrate sensitive systems well before large-scale quantum computers arrive, since data captured today could be decrypted later.

Q

QAOA (Quantum Approximate Optimization Algorithm): A hybrid quantum-classical algorithm designed to find good, though not always perfect, solutions to combinatorial optimization problems such as routing or scheduling. It alternates between a problem-specific quantum operation and a mixing operation, repeated for a chosen number of rounds, while a classical optimizer tunes the parameters between runs. QAOA is one of the most studied candidates for demonstrating practical quantum advantage on optimization problems using near-term, noisy hardware, though its performance depends heavily on how many rounds are used.
QED-C Benchmarks: A benchmarking suite developed by the Quantum Economic Development Consortium that evaluates quantum hardware using circuits tied to real-world domains such as chemistry, finance, and optimization. Rather than relying purely on random circuits, QED-C benchmarks aim to measure performance in a way that predicts practical workload behavior more directly. Results are typically reported per application category, allowing comparison of which platforms suit which problems. The suite is maintained collaboratively across the industry rather than by a single vendor, keeping the comparisons reasonably neutral.
Quantum Advantage: The point at which a quantum computer solves a problem faster, cheaper, or more accurately than any practical classical approach, sometimes called quantum supremacy for the more extreme academic version of this claim. Some demonstrations show an essentially unreachable classical alternative, while others show a modest practical edge where a classical computer could technically solve the same problem, just slower or with far more resources. All confirmed demonstrations to date involve narrow, largely academic problems rather than commercially useful ones.
Quantum Algorithm: A structured sequence of quantum logic gates, and sometimes measurements, designed to solve a specific computational problem. It may be expressed as a single circuit or a collection of circuits combined with classical processing in between, much like classical algorithms can be built from smaller subroutines. Well-known examples include Shor's algorithm for factoring, Grover's algorithm for search, and QAOA for optimization. Designing a good one usually means finding a way to use superposition and interference to favor correct answers over incorrect ones.
Quantum Annealing: An optimization technique that starts a quantum system in an easily prepared superposition of all possible solutions and slowly evolves it so the system settles into the state representing the best, or a very good, solution. It is applied mainly to combinatorial optimization problems in logistics, scheduling, and certain machine learning tasks. Quantum annealing hardware is architecturally different from gate-based quantum computers and is generally not considered capable of universal quantum computation; its value depends heavily on how the target problem maps onto the hardware's connectivity graph.
Quantum Bit (Qubit): The basic unit of information in a quantum computer, playing the role a bit plays in classical computing. Unlike a classical bit, always definitely 0 or 1, a qubit can exist in a superposition of both states at once, described as a combination of basis states with complex-number coefficients called amplitudes. Qubits can be physically implemented in many ways, including trapped ions, superconducting circuits, photons, and neutral atoms, each with its own tradeoffs. A qubit's real power only appears once multiple qubits are entangled together.
Quantum Chemistry: One of the application areas expected to benefit earliest from quantum computing, because classical computers struggle to simulate the quantum mechanical behavior of molecules past a fairly small size. Quantum computers are naturally suited to this task since molecular interactions are themselves governed by quantum mechanics. Promising use cases include drug discovery, catalyst design, better battery materials, and modeling processes like nitrogen fixation for fertilizer production. Algorithms such as the Variational Quantum Eigensolver are among the leading near-term approaches being tested here.
Quantum Circuit: The standard way of representing a quantum computation, drawn as a diagram with one horizontal line per qubit and boxes placed along each line for the gates applied to it in order. A quantum circuit may also include measurement operations, usually shown at the end of each qubit's line, converting the quantum state into a classical readout. Almost every quantum programming framework is built around this circuit model. The order gates appear in the diagram directly matches the sequence they are physically applied to the hardware.
Quantum Cloud Computing: The practice of accessing quantum hardware remotely over the internet rather than owning and operating a quantum computer directly. Major cloud platforms in this space offer access to a mix of proprietary and partner hardware, alongside simulators for cheap local testing before spending time on real devices. This model has been central to opening quantum computing up to researchers, students, and companies who would otherwise have no way to run circuits on real hardware, and it remains the primary way most users interact with quantum processors today.
Quantum Computer Benchmarking: The general practice of running standardized quantum programs on a device to characterize its overall performance in a single, comparable way. It aims to combine lower-level statistics, such as gate error rates, connectivity, and coherence time, into a higher-level score that is easier to compare across machines. Different approaches exist, including randomized benchmarks like Quantum Volume and algorithmic benchmarks tied to real applications, each with different strengths, and no single benchmark fully captures a device's usefulness on its own.
Quantum Error Correction (QEC): A collection of techniques for protecting quantum information from noise and decoherence by encoding it redundantly across multiple physical qubits. Unlike classical error correction, QEC has to work without ever directly reading out the protected state, since measuring it would destroy the information being protected. Instead, QEC codes detect specific error patterns through indirect measurements and correct them without collapsing the underlying computation. The surface code is the most widely pursued approach because it tolerates relatively high physical error rates.
Quantum Error Mitigation: A set of techniques that reduce the impact of noise on a quantum computation's results without the full overhead of formal error correction. Common methods include zero-noise extrapolation, which runs a circuit at several artificially amplified noise levels and extrapolates back to an estimated noise-free result, and probabilistic error cancellation, which statistically compensates for known error patterns. These trade extra classical processing and repeated runs for improved accuracy on today's noisy hardware, ahead of when full error correction becomes practical.
Quantum Gate: A basic operation applied to one or more qubits, serving the role a logic gate serves in a classical circuit. Quantum gates are unitary transformations, meaning they preserve the total probability of a quantum state while changing its details. Some gates act on a single qubit, such as the Hadamard gate, which creates superposition, while others act on pairs of qubits, such as the CNOT gate, which creates entanglement. Chaining gates in a specific order is how a quantum circuit implements a full algorithm.
Quantum Information Science (QIS): An interdisciplinary field studying how information can be stored, processed, and transmitted using quantum mechanical systems, sitting at the intersection of physics, computer science, and mathematics. QIS is broader than quantum computing alone, also encompassing quantum sensing, quantum communication, and quantum networking. Researchers work on everything from the underlying physics of qubits to the algorithms and software that run on top of them, and most academic programs in this space are organized under the QIS umbrella rather than quantum computing specifically.
Quantum Interference: The way probability amplitudes of different quantum states combine when they overlap, either reinforcing each other through constructive interference or canceling out through destructive interference. Quantum algorithms deliberately engineer interference so the amplitude for a correct answer builds up while amplitudes for incorrect answers cancel, making the correct answer far more likely upon measurement. Without interference, superposition alone would offer no computational benefit, since you would just be sampling from a large space of equally likely outcomes.
Quantum Internet: A proposed future network connecting quantum devices across long distances using quantum communication channels rather than purely classical ones. Its core building block is distributing entanglement between distant nodes, enabling applications such as provably secure communication through quantum key distribution and distributed quantum computing across multiple smaller processors. Realizing it at scale depends on technologies still maturing, particularly quantum repeaters and quantum memories capable of holding entangled states long enough to be useful, though early satellite and fiber links have already been demonstrated.
Quantum Key Distribution (QKD): A method for two parties to establish a shared secret encryption key with security guaranteed by the laws of physics rather than the difficulty of a mathematical problem. Protocols such as BB84 exploit the fact that measuring an unknown quantum state necessarily disturbs it, so any eavesdropper attempting to intercept the key leaves a detectable trace. QKD is one of the more commercially mature quantum technologies, with working deployments already operating over metropolitan fiber and some satellite links, paired with conventional encryption for the actual data.
Quantum LDPC Code (qLDPC): A family of quantum error correction codes based on low-density parity-check constructions, drawing on the classical LDPC codes used in modern communication. Their appeal is efficiency: they promise far fewer physical qubits per logical qubit than the surface code, which would sharply cut the hardware cost of fault tolerance. The tradeoff is that they generally require interactions between qubits that are not physically adjacent, which is difficult on hardware limited to nearest-neighbor connectivity, though more natural on platforms such as neutral atoms. Resource estimates using qLDPC codes have substantially lowered projected qubit counts for tasks like integer factoring, making them one of the most active areas of error-correction research.
Quantum Machine Learning: An interdisciplinary field exploring how quantum computing and machine learning can enhance one another. This includes using quantum computers to potentially speed up parts of a classical ML pipeline, using classical ML to help design and optimize quantum circuits, and developing entirely new learning models native to quantum hardware, such as quantum kernel methods. Practical, provable advantages over classical methods on real-world data remain an active and somewhat debated area, with most current work focused on small-scale demonstrations rather than production systems.
Quantum Measurement: The process of extracting classical, readable information from a quantum system, typically performed at the end of a computation. When a qubit in superposition is measured, its state collapses to a definite 0 or 1, with the probability of each outcome determined by the corresponding amplitude squared. Measurement is irreversible, so quantum algorithms are carefully designed around exactly when it happens. Randomness only appears at this final step; up to that point, every stage of a noise-free computation evolves deterministically.
Quantum Mechanics: The branch of physics describing the behavior of matter and energy at the scale of atoms and subatomic particles. Its rules differ sharply from everyday classical intuition, allowing phenomena like superposition, where a system exists in multiple states at once, and entanglement, where particles stay correlated regardless of distance. Quantum computing is an attempt to harness these rules directly for information processing rather than treating them as a curiosity to be worked around. Nearly every term in this glossary ultimately traces back to some property of quantum mechanics.
Quantum Noise: The general term for unwanted disturbances that corrupt a qubit's state during a computation, including bit-flip errors, phase-flip errors, and more general depolarizing noise. It arises from imperfect control hardware, unwanted interaction with the environment, and crosstalk between neighboring qubits on the same chip. As noise accumulates over a circuit, the probability that the final measurement reflects the correct answer steadily decreases, which is why characterizing a device's noise in detail is a prerequisite for effective error mitigation and correction.
Quantum Parallelism: The property that lets a quantum computer effectively evaluate a function across many different inputs at once by applying it to qubits held in superposition. On its own this does not directly hand you a useful answer, since a naive measurement returns one random outcome from all those possibilities. Extracting something useful requires carefully designed interference that makes correct or useful answers far more likely to be observed, which is exactly the trick used by algorithms such as Shor's and Grover's.
Quantum Processing Unit (QPU): The industry term for a complete quantum computing system, including the physical qubits and all supporting hardware needed to control and read them out. For a trapped-ion system this includes the ion trap chip, the trapped ions acting as qubits, the lasers used for gates, and the surrounding electronics and vacuum systems. The term is used deliberately in parallel with CPU and GPU, framing a QPU as another kind of specialized processor a broader computing system can call on for specific tasks.
Quantum Random Number Generator (QRNG): A device that produces genuinely random numbers by measuring an inherently random quantum process, such as detecting individual photons or observing vacuum fluctuations. Unlike classical pseudorandom generators, which are ultimately deterministic algorithms that only look random, a QRNG's output is random as a direct consequence of physics rather than computational complexity. This makes QRNGs particularly valuable in cryptography, and they are already commercially available as standalone hardware, used across gaming, security, and scientific simulation.
Quantum Repeater: A device designed to extend the practical range of quantum communication by overcoming the steady loss of photons as they travel through optical fiber. Rather than amplifying a signal the way a classical repeater does, which is impossible without destroying quantum information, a quantum repeater uses techniques like entanglement swapping to relay quantum states between distant nodes without directly measuring them. Building practical, high-rate quantum repeaters is widely considered one of the hardest remaining challenges on the path to a working quantum internet.
Quantum Roadmap: A public plan published by a quantum computing company or research organization describing its intended technical milestones and rough timelines for reaching them. Examples include commitments to scale qubit counts, targets for improving gate fidelity, and projected dates for reaching various levels of error correction. Roadmaps help investors, partners, and the research community track progress and set expectations. Because the field moves quickly, timelines are frequently revised and are best read as a statement of current priorities rather than a guaranteed schedule.
Quantum Sensing: The use of quantum mechanical effects, particularly superposition and entanglement, to build measurement devices more sensitive than anything achievable with classical instruments. Applications already in commercial or near-commercial use include atomic clocks for precision timing, magnetometers for certain medical imaging, gravimeters for geological surveying, and inertial navigation that does not rely on GPS. Quantum sensing is generally considered the most commercially mature branch of quantum technology, with working products already sold well ahead of large-scale quantum computing.
Quantum Simulation: The use of a controllable quantum system to model the behavior of another quantum system too complex to simulate accurately on a classical computer. This was Richard Feynman's original 1982 motivation for building quantum computers at all, reasoning that nature is fundamentally quantum mechanical and should be simulated with quantum hardware. Near-term applications are expected mainly in chemistry, materials science, and condensed matter physics, where accurately modeling atomic interactions could accelerate discovery of new drugs, materials, and industrial processes.
Quantum Software Development Kit (SDK): A programming toolkit that lets developers build, test, and run quantum circuits without working directly with low-level hardware instructions. These SDKs offer high-level building blocks for constructing circuits and then compiling, or transpiling, them down to whatever a specific quantum backend requires. They typically support running the same circuit on a local simulator first and then, without major changes, on real cloud-hosted quantum hardware. Choosing one is often less about technical differences than which cloud ecosystem a developer already works within.
Quantum Teleportation: A protocol that transfers an unknown quantum state from one location to another using pre-shared entanglement and ordinary classical communication, without physically moving the particle carrying the state. Despite the name, the protocol still requires sending classical information at normal speed to complete the transfer, so it cannot communicate faster than light. It is a foundational building block for future quantum networks and distributed quantum computing, and has already been demonstrated experimentally over distances exceeding a thousand kilometers using satellite links.
Quantum Volume: A single-number benchmark intended to summarize a quantum computer's overall practical capability by combining its qubit count, connectivity, gate error rates, and measurement fidelity into one score. It works by finding the largest square circuit, one with equal width and depth, that a device can run successfully within a defined error threshold. Because it accounts for more than raw qubit count, Quantum Volume allows rougher but more meaningful comparisons across very different hardware architectures, though it does not capture every aspect of real-world performance.
Qubit Count: The simplest and most commonly quoted statistic for describing a quantum computer, but also one of the least informative on its own. Because qubit quality, connectivity, and error rates matter just as much as sheer quantity, a headline qubit count needs to be read alongside other metrics like gate fidelity and coherence time to say anything meaningful. A device with fewer, higher-quality qubits can often outperform one with many more low-quality ones, even though qubit count is the easiest number for marketing materials to lead with.
Qubit Lifetime: How long a physical qubit remains usable as a computational resource, often reported as its T1 time. For synthetic qubit technologies such as superconducting circuits, this is typically microseconds to a few milliseconds, directly limiting how long a computation can run. Trapped-ion qubits are a notable exception, since the qubit is literally an atom rather than an engineered structure, giving coherence times measured in seconds or minutes, limited mainly by how long the ion can be successfully trapped and controlled.

R

Randomized Benchmarking: A family of benchmarking techniques that use randomly generated quantum circuits, rather than application-specific ones, to characterize a quantum device's performance. Because the circuits are random, results are rigorous and easy to compare across different hardware platforms, but they only give a directional sense of real-world performance rather than a direct prediction. Quantum Volume, Gate Set Tomography, and Cross-Entropy Benchmarking are all examples used across the industry, typically alongside, rather than instead of, algorithmic benchmarks.

S

Shor's Algorithm: A quantum algorithm, developed by mathematician Peter Shor in 1994, that can factor large integers exponentially faster than the best known classical algorithms. Its significance comes from cryptography, since widely used public-key systems such as RSA rely on the practical difficulty of factoring large numbers on classical hardware. A sufficiently large, fault-tolerant quantum computer running Shor's algorithm could in principle break this class of encryption, which is the main reason post-quantum cryptography research exists. Running it at cryptographically relevant scale is still well beyond today's hardware.
Single-Qubit Fidelity: A measure of how accurately a quantum processor performs operations on an individual qubit, often abbreviated 1Q fidelity. It is generally the easier fidelity metric to achieve at a high level, since single-qubit gates are physically simpler than the two-qubit gates needed for entanglement. When reading a reported number, check whether it reflects the best qubit measured, an average across the device, or another statistic, since these can paint quite different pictures. High single-qubit fidelity alone is not sufficient, since most algorithms are limited by noisier two-qubit gates.
Square Circuit: A quantum circuit with a width of N qubits and a depth of roughly N squared gates, used as a standardized benchmarking shape. Successfully running a square circuit at a given size is often treated as the practical minimum bar for saying a device can use all its qubits effectively together, rather than just running shallow circuits on a large qubit count. This shape is the basis for the Quantum Volume benchmark, providing a consistent test case across very different hardware architectures.
State Preparation and Measurement (SPAM) Error: The combined error introduced at the very start and end of a quantum computation, when qubits are initialized to a known state and later read out for a result. Unlike gate errors, SPAM error generally does not compound repeatedly within a single algorithm, since most current circuits only initialize and measure once. It sets a hard floor on how accurate any result can be, no matter how good the gates themselves are, and becomes a bigger factor as circuits begin using mid-circuit measurement and reset.
Superconducting Qubit: A qubit built from a specialized superconducting circuit, typically incorporating a nonlinear component called a Josephson junction, operated near absolute zero inside a dilution refrigerator. This platform benefits from fast gate speeds and can be fabricated using techniques adapted from the semiconductor industry. Tradeoffs include sensitivity to fabrication imperfections, comparatively short coherence times, limited native connectivity, and the practical difficulty of scaling a single cryogenic system to house very large qubit counts. The transmon is the dominant variant in use today.
Superposition: A core principle of quantum mechanics stating that a system can exist in a combination of multiple states simultaneously until measured. For a qubit this means representing 0 and 1 at once, though the common description of a qubit being 0 and 1 at the same time only scratches the surface, since superposition allows any weighted combination of the two states, not just a fixed midpoint. It is one of two central ingredients, alongside entanglement, giving quantum computers their potential advantage; on its own it must still be paired with interference to produce useful answers.
Surface Code: A leading quantum error correction code that arranges physical qubits in a two-dimensional grid, with each qubit only needing to interact with its immediate neighbors. It is attractive because it tolerates relatively high physical error rates, around one percent, compared to many alternative codes, making it compatible with error rates current hardware can realistically achieve. Its main downside is the large number of physical qubits required to encode each logical qubit, which is why newer low-density parity-check (qLDPC) codes, aimed at sharply lower overhead, have become an active area of hardware research.

T

Topological Qubit: A theoretical qubit design that encodes quantum information in the global, collective properties of exotic quasiparticles called anyons, rather than in the state of a single localized particle. Because the information is spread out topologically rather than stored in one fragile location, this approach is expected to be inherently more resistant to local noise than other qubit types. If topological qubits can be realized reliably at scale, they could dramatically reduce the number of physical qubits needed per logical qubit compared to other error correction approaches.
Transmon Qubit: The most widely used design of superconducting qubit, developed at Yale University in 2007 to reduce a specific sensitivity to electrical charge noise that limited earlier superconducting designs. It works by operating in a regime where a particular circuit energy term dominates over another, smoothing out the qubit's response to small environmental fluctuations. Its relatively straightforward fabrication combined with reasonably good coherence has made it the default choice across most major superconducting efforts; a generic superconducting qubit today is very often specifically a transmon.
Trapped-Ion Quantum Computing: A quantum computing platform that uses individual ions, electrically charged atoms, held in place inside an ion trap chip and manipulated with precisely controlled laser or microwave pulses. Compared to other platforms, trapped ions offer some of the longest coherence times and highest gate fidelities demonstrated to date, along with natural all-to-all connectivity within a single trap. The main tradeoff is that laser-controlled gates tend to run slower than superconducting counterparts, though longer coherence times largely compensate in practice.
Two-Qubit Fidelity: A measure of how accurately a quantum processor performs entangling operations between pairs of qubits, often abbreviated 2Q fidelity. It is typically the most important single quality metric for a quantum device, since two-qubit gates are both the noisiest operations on almost every hardware platform and the operations that create the entanglement most algorithms depend on. Sustained two-qubit fidelities above roughly 99.9 percent are widely regarded as a rough threshold needed for practical, large-scale fault-tolerant quantum computing.

V

Variational Quantum Eigensolver (VQE): A hybrid quantum-classical algorithm designed to estimate the lowest possible energy state, the ground state, of a quantum system, most commonly a molecule. A quantum computer prepares a parameterized trial state and measures its energy, while a classical optimizer repeatedly adjusts the parameters to search for the lowest value found. VQE is considered one of the most promising near-term algorithms because it is relatively tolerant of hardware noise compared to algorithms that require deep, error-free circuits, with quantum chemistry and materials science as its main proposed applications.

Y

Ytterbium: A silvery rare-earth metal, atomic number 70, used as the ion species in most current commercial trapped-ion systems. It is well suited to this role because its electronic structure can be manipulated using only a small number of distinct laser wavelengths, simplifying the control hardware needed for gates and readout. As a naturally occurring atom rather than an engineered structure, every ytterbium ion is identical to every other, removing the fabrication variability that affects manufactured qubit types like superconducting circuits. Its long coherence time and simple control requirements make it a practical default for ion-trap hardware.

Z

ZCC-v0.1 (ZKSF Convergence Certificate): An open, versioned protocol for stating the accuracy of an approximate quantum simulation, rather than reporting a bare number with no indication of how far it might be from the truth. Its default mode is a convergence check: the circuit is re-run at a higher resource setting, such as a doubled bond dimension, and the shift in the leading outcome probabilities is reported as evidence the approximation has settled. Its certified mode instead returns a rigorous, single-run bound: for a matrix product state built without renormalization, the final state's norm deficit is the exact weight discarded by truncation, and the error on any outcome is bounded by the square root of twice that value. Any finished job can be exported as a public, independently verifiable certificate under this protocol.
ZHF-v0.1 (ZKSF Hardware Fidelity certificate): The hardware counterpart to ZCC-v0.1: an open protocol for stating how closely a real quantum processor's measured output matches the ideal answer, rather than only reporting raw counts or citing a vendor's own device specification. Where a circuit is small enough to also simulate exactly, the certificate reports the measured fidelity between the hardware counts and that exact ideal distribution, computed directly from the run rather than asserted. Where a circuit is too large for direct verification, the certificate says so honestly instead of reporting an unsupported number. Like ZCC-v0.1, any finished hardware job can be exported as a public, independently verifiable certificate.

Want to see these ideas in action? Our SDK documentation walks through real circuits, real hardware runs, and what the error certificates on each result actually mean. Missing a term you think belongs here? Tell us.