Short answer: Every control and readout line that runs into a dilution refrigerator brings heat to the cold stages, so cooling power limits how many superconducting qubits one fridge can serve. A 2019 EPJ Quantum Technology paper describes a thermally optimized wiring setup, a conference abstract from the University of Technology Sydney describes a numerical wiring model, and a 2025 EPJ Quantum Technology paper measures a common coaxial cable and estimates processor limits of roughly 200 qubits in theory and 140 in practice for that architecture.
Why wiring matters
The 2019 paper says a robust cryogenic infrastructure, a wired and thermally optimized dilution refrigerator, is essential for solid-state quantum processors. The Sydney abstract adds that large processors need many control lines, and that these lines provide an open channel for noise and heat to reach the qubits, creating competing demands that attenuation and filters address. It says the balance becomes more delicate as processors approach the spatial and cooling capacities of dilution refrigerators.
2019: a thermally optimized setup
The authors engineer a cryogenic setup that minimizes passive and active heat loads while keeping qubit control and readout fast. They review design criteria for drive lines, flux lines and output lines, and describe each type. They measure the passive heat load of stainless steel and NbTi coaxial cables and the active load from signal dissipation, validating their concept for thermal anchoring of attenuators, cables and other microwave components.
2025: measuring a common cable
The 2025 paper notes that adding lines without prior thermal modelling risks overwhelming the cooling power at each stage. The authors study the thermal load of a semi-rigid coaxial cable commonly used for control and readout lines. They check its materials and dimensions, measure the total thermal conductivity of one cable from cryogenic to room temperature, and measure the cryogenic DC resistance of the inner conductor to compute heating from Ohmic losses. They fit the data to compute static and active heat loads in a cryogenic model of a processor in a commercial dilution refrigerator, and look at how the load grows with processor size. They conclude that the theoretical upper limit of the described architecture is approximately 200 qubits, and that with an engineering margin in cooling power and mixing-chamber space the practical limit is approximately 140 qubits.
Limits
The 140 and 200 qubit figures belong to the authors' specific cable, refrigerator model and wiring architecture, and are not general limits for all designs. The Sydney text available here is only the start of a conference abstract, so its model results are not covered.
FAQ
Why do more qubits mean more heat?
Each additional control or readout line adds passive heat conduction and active Ohmic heating at the cold stages, according to the 2025 paper.
What limit did the 2025 paper find?
About 200 qubits in theory and about 140 in practice for its described architecture and refrigerator model.
Does that apply to every fridge or design?
No. It applies to the paper's cable, fridge model and layout.
How is heat managed in practice?
The 2019 paper uses thermal anchoring of attenuators, cables and other microwave components, and measures stainless steel and NbTi cables.
Does wiring affect noise as well as heat?
The Sydney abstract says control lines also let noise propagate to the qubits, which attenuation and filters address.
All field notes