The advancement of photonic quantum technologies is accelerating their transition from laboratory demonstrations to real-world applications in quantum computing, communication, and metrology. Scaling these systems creates a pressing demand for high-fidelity, massively parallel single-photon detection. The criticality of this capability is underscored by recent frontier experiments. In optical quantum computing, the “Jiuzhang” series manipulated 76 to over 200 photons, a feat supported by more than a hundred high-efficiency single-photon detectors, directly tying computational complexity to available detection resources. Similarly, quantum entanglement networks have expanded to 40 nodes. Further scaling of such networks depends not only on enhanced entanglement sources but is also fundamentally limited by the capacity and performance of parallel detection infrastructures.

Semiconducting detectors, such as single-photon avalanche diodes (SPADs), are technologically mature and widely deployed. Nevertheless, their performance in key metrics, including detection efficiency, timing resolution, and dark count rates, often falls short of the stringent requirements of advanced quantum photonic applications. In contrast, superconducting nanowire single-photon detectors (SNSPDs) represent the current state-of-the-art, offering near-unity detection efficiency, ultra-wide response bandwidth , picosecond-scale timing resolution, GHz-rate operation, and ultralow dark counts.

Despite these advantages, the large-scale deployment of SNSPDs remains challenging. To maintain their superconducting state, SNSPDs must operate at cryogenic temperatures, typically below 2.2 K. However, each detection channel, whether for current biasing or signal readout, requires independent coaxial interconnects across various temperature stages. In a parallelized configuration, the passive heat load introduced by these cables scales linearly with the channel count, eventually exhausting the cooling power of the cryostat and degrading detector performance. Furthermore, integrating a high density of detectors within a confined cryogenic volume presents significant challenges in device packaging, fiber-optic routing, and electrical feedthroughs. Consequently, most practically deployed SNSPD systems have been restricted to approximately 16 channels.

Recent efforts have sought to increase the integration scale, yet each approach involves distinct trade-offs. Chip-level integration using waveguide coupling has reached 64 channels but often suffers from excessive coupling losses and concentrated thermal loads. System-level solutions employing high-power, large-scale cryocoolers also achieve 64 channels but incur substantial penalties in physical footprint and power consumption. Alternatively, specialized interconnects using low-thermal-conductance cables offer a promising route by addressing the thermal bottleneck, though scaling such solutions to the centennial level remains unproven. These approaches, often focused on component-level optimization, fail to balance the competing requirements of detection performance and system scalability. Therefore, realizing high-performance, massively parallel detection requires a transition to holistic system-level co-design, which must simultaneously address these interdependent challenges: guaranteeing high detection performance during large-scale integration; overcoming the interconnect thermal load bottleneck; developing parallel bias and readout electronics, and minimizing the overall cryogenic overhead.

In this work, we transcend these limits through a scalable architecture that systematically integrates detector performance, thermal management, readout electronics, and cooling requirements. As experimental validation, we report a 100-channel platform operating stably below 2.1 K within a compact Gifford-McMahon (GM) cryocooler. By integrating 100 independently selected high-performance SNSPDs, the system achieves system detection efficiencies (SDE) ranging from 80.1% to 98.3%, with dark count rates (DCR) below 300 cps across all channels. This platform represents the largest SNSPD system demonstrated to date. Finally, we showcase its platform utility by implementing a fully connected multi-user quantum entanglement distribution network, providing a practical and robust infrastructure for next-generation quantum technologies.