What Is NISQ Quantum Computing: The Current Era of Quantum Machines

Insider Brief

• NISQ, or Noisy Intermediate-Scale Quantum, describes current quantum computers that can run quantum algorithms but remain limited by noise, error rates and circuit depth.

• NISQ systems are primarily used for research and experimentation, including testing quantum algorithms, error mitigation techniques and hybrid quantum-classical approaches.

• The industry is beginning to move toward fault-tolerant systems through advances in quantum error correction, although NISQ hardware remains the dominant category of commercially available quantum computers.

Every quantum computer commercially available today belongs to the same category. Whether it is IBM ‘s quantum systems, Google ‘s superconducting processors, or IonQ ‘s trapped-ion hardware. All of these are NISQ devices, whether the press releases say so or not.

NISQ stands for Noisy Intermediate-Scale Quantum. Physicist John Preskill coined the term in 2018 to describe machines that are large enough to be difficult to simulate on classical computers, but too error-prone to run complex algorithms that would make quantum computing commercially transformative. The label has become standard across the industry because it accurately describes where quantum hardware is.

Breaking encryption at scale, simulating large molecules accurately, solving optimization problems faster than classical computers at production scale. None of these are within reach of current NISQ devices. They serve as research platforms for testing algorithms, developing error mitigation techniques, and building the expertise that may eventually support more capable systems.

What does NISQ stand for?

Each word in NISQ describes a defining characteristic of current quantum hardware.

Noisy

A qubit is the basic unit of information in a quantum computer, equivalent to the bit in a classical computer. Unlike a classical bit, which holds a fixed 0 or 1, a qubit can exist in a combination of both states at once, a property called superposition. This makes qubits powerful, but it also makes them fragile.

Qubits are extremely sensitive to their surroundings. Stray electromagnetic fields, temperature fluctuations, and small imperfections in the control signals used to manipulate them all introduce errors. Quantum states also degrade over time through a process called decoherence, where the qubit loses its quantum properties through unwanted interactions with the environment. Every operation on a qubit carries a chance of producing an incorrect result, and these errors accumulate as a computation proceeds, limiting how long a calculation can run before the results become unreliable.

Intermediate-Scale

Current systems have qubit counts ranging from tens to a few thousand. This is large enough that simulating these systems on classical computers becomes difficult or impossible, which is what makes them scientifically interesting. It is far smaller than the millions of physical qubits that fault-tolerant quantum computing may require for commercially valuable applications.

Quantum

The Q in NISQ simply confirms what these machines are: computers that use quantum mechanical effects to process information. Superposition allows qubits to represent multiple states simultaneously. Entanglement links the states of separate qubits so that measuring one instantly affects the other. Interference allows quantum algorithms to amplify correct answers and cancel out wrong ones. These are the properties that separate quantum computers from classical ones, and that give the NISQ category its defining characteristic even in this limited, error-prone form.

Together, the three terms describe machines that are real and functional but constrained. They demonstrate genuine quantum behavior. They cannot yet do the things that would make quantum computing commercially transformative.

Who Coined the Term NISQ?

John Preskill, a theoretical physicist at the California Institute of Technology, introduced the term NISQ in a 2018 paper and an accompanying keynote. His motivation was to give the field a realistic way to discuss the machines actually being built, rather than the idealized fault-tolerant systems that dominated theoretical discussions at the time.

Before NISQ, conversations about quantum computing often jumped straight to fault-tolerant machines running algorithms like Shor’s for factoring large numbers. Preskill’s framing acknowledged that the devices actually being built would be limited in many ways. It gave researchers and the industry a shared vocabulary for the gap between early hardware and the systems that would eventually justify quantum computing’s broader ambitions.

Since 2018, NISQ has become standard terminology. Papers discuss NISQ algorithms designed around noisy hardware constraints, and the industry uses the term to set realistic expectations about what current technology can and cannot do.

Why are NISQ Computers Noisy?

Qubits must be isolated from their environment to maintain quantum coherence, but they must also be manipulated and measured, which requires interacting with them. This tension between isolation and control is a central engineering challenge across every hardware platform.

Decoherence

Qubits lose their quantum properties through unwanted interactions with their environment, causing quantum states to degrade over time. The coherence time varies by technology, typically microseconds to milliseconds for superconducting systems. But once a qubit decoheres, its quantum information is gone.

Gate Errors

The operations that manipulate qubits are imperfect. Applying a quantum gate involves precisely controlled pulses of energy, and small inaccuracies in timing, amplitude, or frequency introduce errors. For many current NISQ systems, two-qubit gate error rates fall between roughly 0.1% and 1%. Trapped-ion platforms generally achieve lower error rates than superconducting systems at comparable qubit counts, with some prototypes reaching 99.99% two-qubit gate fidelity as of late 2025 , though these results are on pre-commercial hardware.

Measurement Errors

Reading out the final state of a qubit can produce the wrong result. Measurement is a delicate process, and imperfections in detection introduce inaccuracies that affect every computation.

Crosstalk

Operations on one qubit can unintentionally affect neighboring qubits. And as qubit counts increase and systems grow denser, isolating individual qubits from each other becomes even harder.

Each operation also introduces a small chance of error. As more gates are applied, those errors can accumulate and eventually disrupt the computation. This limits NISQ devices to relatively short circuits with fewer operations.

What can NISQ Computers actually do?

NISQ computers are research and experimentation platforms. They cannot yet solve commercially valuable problems that are beyond the reach of classical computers.

They allow researchers to test quantum algorithms on real hardware under realistic noise conditions, revealing which approaches might scale and which face practical obstacles that idealized simulations miss.

Error mitigation has become a key tool for extending what current hardware can do. Rather than eliminating noise, mitigation uses statistical post-processing to infer what a noiseless result would have been. Techniques including zero-noise extrapolation and probabilistic error cancellation can extend useful circuit depth from hundreds of operations to thousands or even tens of thousands . The overhead grows with circuit size, but the techniques meaningfully extend what NISQ devices can do before full error correction becomes available.

Variational quantum algorithms are designed specifically for NISQ constraints. These hybrid approaches combine short quantum circuits with classical optimization loops. The Variational Quantum Eigensolver estimates molecular energies. The Quantum Approximate Optimization Algorithm targets certain combinatorial optimization problems. Demonstrating advantage over classical methods with these algorithms remains an open challenge.

NISQ systems have also demonstrated what researchers call quantum utility. In June 2023, IBM and UC Berkeley published results in Nature showing a 127-qubit Eagle processor produced accurate results for a physics simulation at a scale beyond what classical computers could reliably verify using brute-force methods. IBM described this as the beginning of the “quantum utility” era – NISQ devices producing results of genuine scientific value ahead of fault tolerance. Similarly, in October 2025, Google demonstrated verifiable quantum advantage on a physics simulation problem using its 105-qubit Willow processor .

For a broader view of where practical applications stand, TQI’s guide to quantum computing use cases covers the field across industries.

What are the Limitations of NISQ Computing?

The limitations of NISQ devices follow directly from noise and scale, and they constrain the depth and type of computations these systems can reliably perform.

Limited Circuit Depth

Errors accumulate with each operation. After a certain number of sequential gates, noise overwhelms the computation. NISQ devices are restricted to algorithms that can extract useful results within a limited number of steps, ruling out many algorithms that require deep circuits.

No Full Error Correction

Quantum error correction encodes information across many physical qubits to detect and correct errors, but it requires large numbers of high-quality qubits and very low error rates. NISQ devices lack the scale and fidelity to implement full error correction. This is the defining limitation separating NISQ from fault-tolerant quantum computing.

TQI’s guide to quantum error correction covers the techniques and the companies building toward it.

Difficulty Proving Quantum Advantage

While NISQ devices have demonstrated advantage on narrow problems including physics simulations and random circuit sampling, showing advantage on practically valuable computational problems has proven difficult. Classical algorithms and hardware continue to improve, and classical methods often match or exceed what NISQ devices can achieve on real-world tasks.

Limited Qubit Connectivity

In many NISQ systems, qubits can only interact directly with neighboring qubits. Algorithms requiring interactions between distant qubits need additional routing operations, which adds errors and consumes limited circuit depth.

Measurement Constraints

A measurement collapses the quantum state, yielding only a small amount of classical information per run. Algorithms typically require many repetitions to build up statistics, which takes time and is affected by noise throughout.

How does NISQ differ from Fault-Tolerant Quantum Computing?

NISQ devices operate without full error correction, so noise limits their computations to short algorithms. Fault-tolerant quantum computers would use quantum error correction to detect and correct errors continuously, allowing computations of arbitrary length.

Fault tolerance requires encoding logical qubits, the qubits that perform computation, across many physical qubits. Current estimates suggest hundreds to thousands of physical qubits may be needed per logical qubit, depending on the error correction code and physical error rates. A fault-tolerant system capable of transformative applications might require millions of physical qubits.

TQI’s quantum computing hardware landscape overview covers how different hardware approaches are positioned for this transition.

Characteristic NISQ Systems (Today) Fault-Tolerant Systems (Future) Qubit count Tens to low thousands of physical qubits Millions of physical qubits for thousands of logical qubits Error rates ~0.1% to 1% per gate Effectively corrected below threshold Error correction None or partial Full quantum error correction Circuit depth Limited before errors dominate Long computations possible Applications Research, experimentation, narrow demonstrations Cryptography, large-scale simulation, broad use cases Availability Available now Industry roadmaps target 2028 to 2033

Is the NISQ Era Ending?

The NISQ era is not over, but the transition has started becoming visible at the hardware level.

In December 2024, Google’s Willow chip demonstrated below-threshold error correction , confirming for the first time on real hardware that adding more physical qubits to an error correction code reduces logical error rates.

In June 2025, Quantinuum reported the first universal, fully fault-tolerant quantum gate set with repeatable error correction , describing itself as the first company to step from the NISQ era toward utility-scale quantum computing. The results showed logical error rates surpassing those of physical gates, validating magic state distillation and fault-tolerant non-Clifford gate execution. In March 2026, Quantinuum followed with 94 error-protected logical qubits performing beyond-break-even operations.

Across the industry, roadmaps are converging on the late 2020s. IBM targets fault-tolerant performance by 2029. IQM has outlined a transition from NISQ to error-correction processors by 2027, with early fault-tolerant performance expected in 2028. Oxford Quantum Circuits is targeting a 200-logical-qubit system in 2028. Quantinuum’s Apollo system is planned for 2029.

A Riverlane survey of more than 300 quantum professionals conducted in 2025 found 2028 emerging as an informal industry deadline, with 95% of respondents rating quantum error correction as essential for scaling the technology.

The transition is likely gradual rather than a clean break. Early fault-tolerant systems with small numbers of logical qubits may operate alongside larger NISQ devices. The NISQ label may persist in some form even as the field advances beyond its strictest definition.

For readers looking to go deeper, TQI’s understanding of the quantum computing hardware landscape covers each hardware modality in depth; understanding quantum error correction covers the techniques and vendors building toward fault tolerance; and quantum computing use cases and real applications in 2026 maps where practical applications currently stand.

Frequently Asked Questions

What does NISQ stand for?

NISQ stands for Noisy Intermediate-Scale Quantum. The term describes quantum computers that are noisy (prone to errors), intermediate in scale (tens to a few thousand qubits), and genuinely quantum. Physicist John Preskill coined it in 2018.

Are all current quantum computers NISQ devices?

Yes. Every quantum computer commercially available today, including systems from IBM , Google , IonQ , Rigetti, and others, falls into the NISQ category. These machines lack the full error correction needed for fault-tolerant quantum computing. Quantinuum claimed to cross from NISQ territory in June 2025, but fully capable fault-tolerant systems have not yet been deployed at commercial scale.

What can NISQ computers be used for?

NISQ computers are used primarily for research and experimentation: testing quantum algorithms on real hardware, developing error mitigation techniques, and exploring hybrid quantum-classical algorithms like the Variational Quantum Eigensolver. They cannot yet solve commercially valuable problems better than classical computers on general tasks.

Why are NISQ computers noisy?

Qubits are extremely sensitive to their environment. Decoherence causes quantum states to degrade over time. Gate operations introduce errors at rates between roughly 0.1% and 1% per operation for most current systems. Measurement can produce incorrect results. Crosstalk between neighboring qubits adds further errors. These sources accumulate as computations proceed.

What is the difference between NISQ and fault-tolerant quantum computing?

NISQ devices operate without full error correction, limiting computations to short algorithms before noise overwhelms results. Fault-tolerant quantum computers would use quantum error correction to detect and correct errors continuously, allowing long computations. Fault tolerance requires encoding logical qubits across many physical qubits, potentially millions, far beyond current systems.

When will the NISQ era end?

The transition is already beginning. Google demonstrated below-threshold error correction in December 2024. Quantinuum reported a universal fault-tolerant gate set in June 2025 and 94 logical qubits in March 2026. IBM , IQM, Oxford Quantum Circuits, and Quantinuum all target fault-tolerant systems between 2028 and 2029. The transition will be gradual rather than a single event