Physical Qubits vs. Logical Qubits: Why the Headline Number Rarely Tells the Whole Story

When a quantum computing company announces a new chip with thousands of qubits, that number is usually the one that lands in press releases and investor decks. It is also, in most cases, the least useful number for understanding what the machine can actually do. To make sense of it, you need to understand the difference between physical qubits and logical qubits, and why that gap matters differently depending on the type of quantum hardware you are working with.

What a Physical Qubit Is, and Why It Is Not Enough

A physical qubit is the actual hardware component — the real, physical object that holds quantum information. Depending on the architecture, that might be a superconducting loop, a trapped ion, or a single photon. Physical qubits are capable of superposition and entanglement, which is what makes quantum computation interesting in the first place.

The problem is that physical qubits are extraordinarily fragile. They are sensitive to temperature fluctuations, electromagnetic interference, and even physical vibration. This sensitivity means they have high error rates. Run a calculation long enough on raw physical qubits and the quantum state collapses into noise before you reach a usable answer. The physics is real; but the output is not. Therefore, on its own, physical qubits cannot be used on their own, the same way that transistors are used in silicon-based CPUs. 

What a Logical Qubit Is

A logical qubit is a virtual qubit — not a single physical object, but a collection of physical qubits working together to simulate a stable one. The mechanism that holds them together is Quantum Error Correction (QEC). By spreading quantum information across multiple physical qubits, the system can detect when one of them flips or degrades and correct the error on the fly. The result is a qubit that is stable enough to persist through a long, complex computation.

For the hardware nerds out there – think of this like a RAID array in traditional computing. No single drive is reliable enough on its own, so you distribute the data across several and let the system reconstruct anything that goes wrong. A logical qubit does the same thing with multiple physical qubits.

The cost of this reliability is significant, however, and is the reason why quantum computing remains in the research stage. Depending on the error correction scheme and the quality of the underlying hardware, creating one logical qubit can require dozens or even hundreds of physical qubits. So when a company says it has 1,000 physical qubits, the number of fault-tolerant logical qubits it can actually run may be far smaller… or at times, still effectively zero. It depends on whether error rates are yet low enough to make correction worthwhile.