frequency crowding (collisions)

Frequency crowding is what happens when you try to fit too many qubits into too little radio spectrum. Each fixed-frequency superconducting qubit has its own microwave tone, a few gigahertz, that you use to address it — much like every radio station needs its own slot on the dial so it doesn't bleed into its neighbor. But all the qubits on a chip have to share one finite band, and the band is only a few hundred megahertz wide once you account for the hardware. Pack in a handful of qubits and there is room; pack in a thousand and their assigned frequencies inevitably start landing on top of each other.

A collision is when two frequencies that should be distinct end up too close, and there are several flavors. Two neighboring qubits can share almost the same frequency, so a pulse meant for one also tickles the other. A qubit's frequency can land on a neighbor's, or on the gap a two-qubit gate relies on. Worst of all, a qubit is not a perfect two-level system: its 0-to-1 transition sits near its 1-to-2 transition, and if one qubit's 0-to-1 lines up with another's 1-to-2, gates leak population into states that should never be touched. Each near-coincidence shows up as a gate that misfires, and on a crowded chip these add up fast.

This is one of the genuine walls in front of scaling fixed-frequency designs, not a detail. Because junctions come out of fabrication with scattered frequencies, you cannot simply assign a clean grid of tones and trust the chip to obey — you get what you get, then hope few of them collide. Tunable couplers and flux-tunable qubits dodge the problem by letting you move frequencies after fabrication, at the cost of extra wiring and new noise channels. Better junction targeting shrinks the spread. Neither has made crowding go away; it remains a live constraint that shapes how big a single chip can sensibly get.