Hardware Roadmaps and Scaling
Scaling from tens of qubits to the millions needed for useful error correction is a systems problem spanning qubits, wiring, control, and fabrication yield.
The Gap to Close
Today's processors hold from tens to a few thousand physical qubits. Running a large fault-tolerant algorithm is widely estimated to need on the order of millions of physical qubits, because each logical qubit is encoded across many physical ones and errors must be corrected continuously. Closing this gap is not a matter of any single breakthrough but of advancing every subsystem at once while keeping error rates below the fault-tolerance threshold.
The Coupled Constraints
- Gate and readout errors must stay below the code threshold as the device grows.
- Wiring and cooling cannot scale linearly with qubit count, forcing multiplexing and integration.
- Fabrication yield and frequency uniformity must hold across ever-larger chips.
- Correlated errors from radiation set a ceiling that more qubits alone do not fix.
The Building Blocks
The scaling toolkit runs through this library: tunable couplers to control crosstalk, multiplexed readout and wideband amplifiers to relieve output wiring, flip-chip and through-silicon-via integration to route signals vertically, cryogenic control electronics to cut the room-temperature cable count, and materials and fabrication advances to raise coherence and yield. Modular architectures, linking smaller processors with quantum interconnects, offer a path when a single monolithic chip becomes impractical.
Roadmaps as Coordination
Published hardware roadmaps typically lay out increasing qubit counts alongside targets for error rate, connectivity, and demonstrated error correction, because raw qubit count without falling error rates is not progress. A credible roadmap treats scaling as a co-design problem: qubits, control, wiring, cooling, and software must advance together, and a weakness in any one caps the whole system.
An Honest Assessment
Large-scale, fault-tolerant quantum computing remains a research and engineering goal, not a delivered capability. Steady progress has been made on coherence, gate fidelity, and integration, and error correction has been demonstrated at small scale, but the full path requires sustained advances across every subsystem described here. Understanding the roadmap means understanding that no single number, least of all qubit count, tells the whole story.