QuEra Computing plans to pack over 20,000 physical qubits into a single processing core of its next computer, a substantial increase in density designed to unlock larger workloads. The company detailed its roadmap for a system targeting one billion reliable logical operations, roughly a thousandfold leap beyond the one million operations expected from its current Libra system. Projected specifications also include a 10⁻⁹ logical error rate, addressing a critical challenge in realizing practical quantum computation. Following its roadmap webinar, QuEra detailed a system designed for more than one billion reliable logical operations and is inviting organizations to co-design fault-tolerant applications through the FTQC Founders Circle. “Libra brings fault tolerance to the cloud in 2025, and the next generation is about scaling it by orders of magnitude to unlock new solutions to pressing industry problems,” said Andy Ory, CEO of QuEra Computing. The company is now inviting collaborators to co-design applications for this future hardware.
Gigaquop-Class System Design for One Billion Logical Operations
This scaling is not merely about increasing qubit count; it’s about achieving a level of performance where larger, fault-tolerant workloads become feasible for problems exceeding the capabilities of classical computers. QuEra’s roadmap extends beyond its existing 256-qubit analog computer, Aquila, and the recently unveiled Libra, which is expected to be available on Amazon Braket. The gigaquop system, targeted for initial use from 2026, aims for a projected specification of more than 1,000 logical qubits and a 10⁻⁹ logical error rate, a significant reduction in errors crucial for reliable computation. Achieving this performance necessitates advancements in reducing both space and time overhead, alongside accelerated quantum error-correction decoding. The company is actively pursuing a flexible approach to quantum error correction, moving beyond a single model. QuEra’s neutral-atom platform allows for the exploration and combination of multiple QEC code families tailored to different architectural roles, including memory and operations.
Recent research indicates ultra-high-rate qLDPC code families with an encoding rate approaching 50 percent, potentially reducing the physical-qubit requirements for gigaquop-class machines. QuEra is collaborating with NVIDIA to integrate its quantum processors with NVIDIA’s platform for quantum-GPU supercomputing, aiming to accelerate error correction and qubit calibration. “Building logical qubits at scale requires supercomputers integrating high-performance quantum processors with accelerated computing for tasks such as quantum error correction and qubit calibration,” said Timothy Costa, Vice President and General Manager for Quantum at NVIDIA.
Building logical qubits at scale requires supercomputers integrating high-performance quantum processors with state-of-the-art accelerated computing for tasks such as quantum error correction and qubit calibration.
Timothy Costa, Vice President and General Manager for Quantum at NVIDIA
