Quantum computing startup Alice & Bob has put forward a new design approach aimed at one of the field's stubborn engineering bottlenecks: the tiny delays inside the loops that control superconducting qubits.
According to the Quantum Computing Report, the company is proposing what it calls "decoupled AI topologies" to resolve microsecond control-loop latencies for its superconducting cat qubits.
To unpack that: quantum computers don't just run a calculation and walk away. They constantly measure their qubits and feed corrections back in, because qubits are fragile and lose their quantum state quickly. That measure-and-correct cycle is the "control loop." If the loop is too slow — even by microseconds, millionths of a second — errors can creep in faster than the system can fix them.
Alice & Bob works with "cat qubits," a design named after Schrödinger's famous thought experiment. The pitch behind cat qubits is that they are built to resist certain kinds of errors at the hardware level, which can reduce the heavy overhead other approaches need for error correction.
The Quantum Computing Report frames the company's proposal as restructuring how the control electronics are arranged — decoupling parts of the system and leaning on AI-driven topologies — so the feedback can keep pace with the qubits rather than lagging behind them.
The source item is an announcement of a proposed approach rather than a finished, benchmarked product, so independent performance figures are not provided here.
Why it matters: the race to build useful quantum computers is increasingly being won or lost on unglamorous engineering details like timing and control, not just on adding more qubits. Shaving microseconds off the control loop is the kind of plumbing work that determines whether these machines ever become reliable enough to do real jobs.