A physical qubit is the real, hardware-level quantum bit implemented in a quantum processor — a transmon circuit, a trapped ion, a neutral atom, or any other physical system that stores and processes quantum information. Physical qubits are imperfect: they suffer from decoherence, gate errors, measurement errors, and crosstalk with neighboring qubits. Current physical qubit error rates range from roughly 0.1% to 1% per gate operation, depending on the platform and gate type.

The distinction between physical and logical qubits is central to understanding quantum computing roadmaps. When companies like IBM announce a 1,000-qubit processor, they are counting physical qubits. The number of logical (error-corrected) qubits that can be encoded is far smaller — typically 10-100x fewer, depending on the error correction code and the physical error rate. For example, encoding a single logical qubit with a surface code at current error rates may require 1,000 or more physical qubits.

The physical qubit count of quantum processors has grown rapidly: from 5 (IBM, 2016) to 72 (Google Bristlecone, 2018) to 127 (IBM Eagle, 2021) to 1,121 (IBM Condor, 2023). However, raw qubit count is a misleading metric without considering qubit quality. A 100-qubit processor with 99.9% two-qubit gate fidelity is far more useful than a 1,000-qubit processor with 99% fidelity, because error accumulation in circuits scales exponentially with error rate. The industry is increasingly focusing on qubit quality metrics alongside quantity.