Inside the towering cylindrical refrigerators that cool quantum processors to near absolute zero, a surprisingly mundane problem is holding back progress: too many cables. Every superconducting qubit typically needs its own dedicated microwave line running from room-temperature control electronics down to the millikelvin chip, and as qubit counts push toward the thousands, the resulting tangle of wiring is bumping up against the physical limits of the hardware that keeps everything cold.
A team at Chalmers University of Technology in Sweden has proposed a way around this bottleneck. In a study published in PRX Quantum in early 2026, the researchers describe a time-multiplexing scheme that lets multiple qubits share fewer control cables by routing signals through fast microwave switches placed right next to the processor inside the cryostat. Their simulations, which model systems of up to roughly 1,000 qubits, suggest the approach could slash cable counts without crippling performance.
The wiring wall
The problem is straightforward in principle. A dilution refrigerator can only accommodate so many physical cables before the heat they conduct overwhelms the system’s cooling capacity. At 50 or 100 qubits, the wiring is manageable. At 1,000 or 10,000, it becomes a serious engineering constraint, one that has nothing to do with the quality of the qubits themselves.
“The cabling infrastructure is one of the most underappreciated barriers to scaling,” noted a Chalmers summary distributed through Phys.org in April 2026. The university highlighted a 121-qubit grid, arranged in an 11-by-11 layout, as a key test case in the study.
How the scheme works
Rather than running one cable per qubit, the Chalmers design groups multiple qubit-control targets onto shared lines. Cryogenic microwave switches sitting beside the processor rapidly route time-sliced pulses to the correct qubit at the correct moment, replacing the brute-force one-to-one wiring model that dominates current architectures.
The central finding is that two-qubit coupler drives, the signals that entangle pairs of qubits, can share cables with zero additional runtime cost. For single-qubit drives, the time penalty grows only logarithmically as the system scales. That is a far gentler cost curve than the linear cable growth today’s designs demand, and it means the wiring savings do not come at the price of dramatically slower computation.
The switch hardware underpinning this idea draws on years of device-level research. Earlier studies have demonstrated superconducting switches capable of on-chip routing of quantum microwave fields with low insertion loss at millikelvin temperatures. Separate work has shown cryogenic microwave switches actuated by high-energy electron injection, documenting trade-offs in speed, reset behavior, and magnetic-field tolerance. These prototypes were not originally built for large-scale control fan-out, but their measured performance aligns closely with what the Chalmers multiplexing scheme requires.
A complementary approach, published in npj Quantum Information, takes a different route: generating and multiplexing microwave control signals using adiabatic superconductor logic with ultra-low power dissipation. That work confirms the broader principle that multiplexing strategies can materially reduce the wiring running from room temperature to the coldest stage of a refrigerator, though no single experiment has yet combined both approaches into a unified architecture.
What has not been proven yet
The strongest results so far rest on simulation, not on a working multi-qubit demonstration. No publicly available data shows the time-multiplexing scheme operating on an actual processor with dozens or hundreds of qubits. Individual switch prototypes have been characterized in the lab, but integrating many of them inside a cryostat alongside a live quantum processor introduces thermal, electromagnetic, and timing challenges that models alone cannot fully capture.
One open question is heat. Every switch dissipates a small amount of energy, and at millikelvin temperatures, even microwatts matter. The cumulative thermal load of dozens or hundreds of switches on a dilution refrigerator’s cooling budget has not been measured in a real system. Quantitative dissipation and switching-speed data are available through journal abstracts and preprints, but full supplementary materials from the PRX Quantum and npj Quantum Information papers were not directly accessible for this analysis.
Crosstalk is another concern. When time-sliced pulses share a cable, small timing errors or signal leakage between slots could degrade qubit fidelity. The Chalmers simulations assume idealized switch behavior, and it remains unclear how sensitive the scheme is to fabrication variability, where small deviations in switch characteristics could translate into amplitude errors or timing skew.
Perhaps the most consequential unknown is compatibility with quantum error correction. Large-scale error-correcting codes impose rigid timing structures and demand high parallelism. The PRX Quantum analysis assumes gates can be rearranged to fit into multiplexed time slots with only logarithmic overhead, but whether that flexibility survives the strict scheduling demands of codes like the surface code has not been demonstrated.
Where this fits in the scaling race
The Chalmers work arrives at a moment when the quantum computing industry is grappling openly with infrastructure constraints. IBM, Google, and other groups building superconducting processors have acknowledged that scaling beyond a few hundred qubits will require rethinking not just chip design but the entire stack of supporting hardware, from control electronics to refrigeration. A scheme that reduces cable counts by a large factor without proportionally increasing runtime would address one of the most concrete obstacles on that path.
The PRX Quantum paper, having undergone peer review at the American Physical Society, carries more weight than preprint-only work. Its analytical proofs and numerical simulations provide a credible theoretical foundation. The device-level papers on cryogenic switches confirm that the required components are not purely hypothetical: prototypes exist and have been measured, even if they have not been deployed in the exact configuration the Chalmers team envisions.
Still, the gap between simulation and hardware demonstration is real. The 1,000-qubit and 121-qubit figures describe modeled layouts, not devices currently running in a laboratory. Readers should treat those numbers as indicators of potential scalability rather than descriptions of existing machines.
The evidence base justifies cautious optimism. Time-multiplexed control can, in principle, tame the wiring explosion threatening superconducting quantum computers, and existing switch prototypes show the key components can perform adequately at the required temperatures. But until a research group builds a working demonstration at meaningful scale, this remains a promising blueprint, not yet a proven solution.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
