Mitigating Capacitive Loading Enables Fast Two-Qubit Gates in Highly Connected Fluxonium Quantum Processors
Fluxonium qubits combine long coherence times with strong anharmonicity, making them attractive for scalable superconducting quantum processors. Recent experiments have demonstrated high-fidelity two-qubit gates and multi-qubit entanglement in one-dimensional and connectivity-four fluxonium processors, motivating their extension to highly connected two-dimensional (2D) architectures. A central challenge in this extension is to achieve strong coupling to multiple circuit elements while preserving the finite capacitance budget of fluxonium qubits, which constrains the attainable qubit-qubit effective coupling strength and, consequently, two-qubit gate performance. Here, we establish a unified theoretical framework that identifies the connectivity-dependent upper bound on capacitive coupling and quantifies its reduction by distinct parasitic-capacitance channels, applicable across both fluxonium and transmon regimes. We show that pad-to-ground loading is primarily geometry limited and can be substantially suppressed through qubit-pad engineering, whereas inter-pad loading is ultimately constrained by parasitic capacitances of Josephson junctions and Josephson junction arrays. Building on these insights, we formulate practical design principles and numerically demonstrate two-qubit gates as fast as 27 ns and leakage-induced infidelity below $10^{-3}$ under representative fabrication variations in 2D fluxonium architectures. These results establish a quantitative framework for understanding and mitigating capacitive loading, suggesting that it does not impose a fundamental performance limit on highly connected fluxonium processors.