Research Hub

Demonstrates that increasing surface code distance on the 105-qubit Willow processor reduces logical error rates exponentially, crossing the critical below-threshold milestone for the first time on a superconducting platform. This result validates the fundamental premise of quantum error correction.

Demonstrates 12 logical qubits with real-time active error correction on the H2 trapped-ion processor. Uses the [[7,1,3]] Steane code with fault-tolerant syndrome extraction and achieves logical error rates below 1e-4 per round.

Reports experimental evidence for topological qubit operation in the Majorana 1 chip, using InAs-Al semiconductor-superconductor heterostructures. The device shows signatures consistent with non-Abelian anyonic statistics in a controlled experimental setting.

Demonstrates 48 logical qubits and hundreds of entangling operations between them using a reconfigurable neutral-atom array. The architecture leverages atom shuttling for non-local connectivity and shows color code and surface code implementations.

Demonstrates that a 127-qubit quantum processor with error mitigation can produce accurate results for a condensed-matter physics simulation that cannot be reliably computed by brute-force classical methods. Establishes that noisy quantum computers can be useful before full fault tolerance.

Demonstrates factoring of a 50-bit RSA integer on a trapped-ion quantum processor using an optimized version of Shor's algorithm. While still far from cryptographically relevant sizes, this represents the largest quantum factoring result to date.

Applies quantum approximate optimization algorithms (QAOA) on a 196-qubit neutral-atom processor to real-world combinatorial optimization problems from logistics and finance. Shows that native graph connectivity of neutral atoms provides advantages for certain problem classes.

Demonstrates quantum computational advantage using a programmable photonic processor with 216 squeezed-state modes. The Gaussian boson sampling task is verified to be classically intractable through rigorous complexity-theoretic analysis.

Introduces an AI-assisted quantum circuit transpiler that uses reinforcement learning to discover better gate decompositions and qubit routing strategies. Integrated into Qiskit 2.0, it reduces circuit depth by 30-50% compared to heuristic methods.

Proposes and validates the "algorithmic qubits" metric as a practical measure of quantum computing capability, accounting for qubit count, gate fidelity, connectivity, and measurement accuracy in a single number tied to algorithm performance.