What a Quantum Computer Actually Is
A quantum computer is a physical device that exploits quantum mechanical phenomena directly in its hardware. The qubits are physical objects, realized through superconducting circuits, trapped ions, neutral atoms, photons, silicon spin states, or other quantum systems, whose superposition and entanglement are inherent in their physics rather than implemented in software.
These physical qubits exist in quantum superposition and entanglement because quantum mechanics governs their behavior at the hardware level. When you apply a gate operation to a superconducting qubit, you are sending a precisely calibrated microwave pulse that physically rotates the qubit's quantum state. The entanglement between qubits is a physical relationship among real quantum systems that cannot generally be described or reproduced classically without resources that grow exponentially with system size.
This is both the power and the fragility of physical quantum computers. The quantum behavior is not implemented. It is inherent. Nature performs the computation by evolving physical quantum states rather than by an explicit classical representation of those states.
A fifty-qubit quantum system can be described mathematically by a state vector containing 2^50 amplitudes, more than one quadrillion values. A classical computer can in principle simulate such a system, but the memory and computational costs grow exponentially. A physical quantum computer does not need to store those amplitudes explicitly in classical memory. The quantum state evolves according to the laws of quantum mechanics within the hardware itself.
The price of this power is extraordinary fragility. Quantum states are disrupted by heat, electromagnetic interference, vibration, and even the act of observation. Maintaining coherence across dozens or hundreds of physical qubits requires millikelvin temperatures, heroic engineering, and error correction schemes that consume most of the available qubits simply to protect the ones performing useful work.
A quantum computer capable of running Shor's algorithm against real-world encryption would likely require millions of physical qubits to encode the thousands of reliable logical qubits the algorithm needs. That machine does not yet exist.
What a Quantum Simulator Is
A quantum simulator is a classical computer running software that correctly implements the mathematics of quantum mechanics. It is not a quantum device. It is a program.
The most common form is a state vector simulator. It allocates a vector of 2n complex numbers in memory, one for each possible computational basis state of an n-qubit system, and applies gate operations as matrix multiplications that transform that vector according to the rules of quantum mechanics. When you run a quantum circuit on a state vector simulator, you get exactly the right answer because the mathematics is exact. The simulator is not approximating quantum behavior. It is computing it precisely.
This is genuinely useful. State vector simulators are how researchers develop and test quantum algorithms before running them on real hardware. They are how students learn quantum computing. They are how errors in quantum circuit design are caught before expensive hardware time is consumed. Frameworks such as Qiskit and Cirq include simulators that can produce results indistinguishable from ideal quantum hardware.
But there is a cost that reveals the fundamental distinction. A state vector simulator requires memory proportional to 2^n to represent n qubits. 30 qubits requires about 16 gigabytes. 40 qubits requires 16 terabytes. 50 qubits requires 18 petabytes. The exponential growth of the Hilbert space, the mathematical space of all possible quantum states, must be represented explicitly in classical memory because there is no physical quantum substrate carrying that information implicitly. The simulator eventually hits a memory wall that cannot be eliminated by faster processors alone, because the underlying challenge is mathematical: representing a generic quantum state requires resources that grow exponentially with the number of qubits.
This is the wall that physical quantum computers do not encounter in the same way because they do not explicitly store the full state vector in classical memory. The quantum state evolves physically within the hardware itself.
Why the Distinction Matters
The practical consequence of this distinction is significant and frequently misunderstood in both directions.
A quantum simulator can produce perfectly correct quantum results, including Bell inequality violations, interference patterns, and accurate algorithm outputs, up to the memory limit of the classical hardware running it. Below that limit, its outputs are indistinguishable from a real quantum computer running the same circuit without errors. This makes simulators invaluable for research and development.
But a quantum simulator is not a quantum computer. It cannot scale indefinitely beyond the memory wall. It does not achieve the same path to large-scale quantum advantage because it pays the cost of explicitly representing the quantum state in classical memory.
Running Grover's algorithm on a simulator finds the correct answer, but it does so by representing all 2^n amplitudes and manipulating them classically. A physical quantum computer finds the answer by evolving physical quantum states in superposition according to quantum mechanics, without requiring a classical system to store the complete state vector explicitly.
The confusion between the two is not merely academic. Commercial claims about quantum advantage sometimes blur this line. A system that produces quantum results on small problem sizes may be a very good simulator rather than a device that will scale to the problem sizes where quantum advantage actually matters. Evaluating such claims requires asking not just whether the outputs are correct but whether the hardware achieving them remains subject to the exponential memory constraint of classical simulation or is leveraging physical quantum phenomena directly.
A Note on Advanced Classical Simulators
The distinction becomes even more interesting because not all quantum simulators are state vector simulators.
Specialized techniques such as tensor networks, matrix product states, stabilizer methods, and Clifford simulators can efficiently model certain classes of quantum circuits with hundreds or even thousands of qubits by exploiting the structure in those circuits. These are classical algorithms that borrow mathematical structures from quantum computing. They do not use quantum hardware. They run entirely on classical computers.
These approaches are powerful and continue to advance rapidly. However, for generic quantum circuits with unrestricted entanglement, the computational resources required for classical simulation still grow exponentially with system size. The existence of these specialized simulators does not erase the distinction between simulation and physical quantum computation. It simply means that some quantum systems are easier to simulate than others.
Quantum-inspired algorithms can deliver impressive speedups for specific problem types. They are useful and interesting classical algorithms, but they do not produce genuine quantum behavior and should not be confused with quantum computing itself. Understanding what they are prevents misplaced expectations about what they can and cannot do.
The Line That Matters
The line between a quantum computer and a quantum simulator is ultimately the line between physics and mathematics. A quantum computer exploits physical quantum phenomena inherent in its hardware to escape the exponential cost of representing quantum states classically. A quantum simulator implements quantum mathematics accurately on classical hardware, producing exact results up to the memory limit imposed by the mathematics.
Both are valuable. Both produce genuine quantum results within their respective operating ranges. But only one of them scales to the problem sizes where quantum computing changes what is computationally possible. Understanding which one you are looking at when evaluating hardware, software, or commercial claims is one of the most useful things you can know about the field.