IBM’s qLDPC roadmap: IBM’s new architecture uses low‑density parity‑check codes that cut qubit overhead for error correction by ~90 %, paving the way for modular machines like Starling (200 logical qubits) and Blue Jay  .

Google Willow: Willow scales error‑correction below threshold, halving errors as more qubits are added; it ran a random‑circuit sampling task in five minutes that would take a supercomputer 10^25 years  .

Reliable logical qubits: Quantinuum and Microsoft built four logical qubits with error rates 800× lower than the physical qubits, then entangled 12 logical qubits in a GHZ state with circuit errors of 0.0011  .

50‑qubit entanglement: Quantinuum later entangled 50 logical qubits with >98 % fidelity and used single‑shot error‑correction techniques.

Infleqtion: Neutral‑atom pioneer Infleqtion announced a 1 600‑qubit array and aims to deliver a 100‑logical‑qubit, million‑depth circuit machine within five years  .

IonQ’s scale-up: IonQ plans 100 qubits in 2025, 10 000 on one chip by 2027, and 2 million by 2030, translating to tens of thousands of logical qubits with error rates <10^‑12  .

Algorithmic fault tolerance: Harvard, Yale and QuEra researchers unveiled a framework that combines transversal operations with correlated decoding, promising 10–100× faster execution of large‑scale algorithms  .

Exponential speedup: A USC‑led team used IBM’s 127‑qubit chips to demonstrate an unconditional exponential speedup on a variation of Simon’s problem by combining shorter circuits, dynamical decoupling and error‑mitigation  .

Industrial silicon qubits: Imec and Diraq showed that CMOS‑fabbed silicon quantum‑dot qubits achieve >99 % two‑qubit gate fidelity and >99.9 % SPAM fidelity, making high‑volume manufacturing feasible .

Qudit error correction: Yale researchers demonstrated error‑corrected qutrits and ququarts using the GKP bosonic code, pushing these higher‑dimensional qudits past break‑even

IBM’s qLDPC roadmap: IBM’s new architecture uses low‑density parity‑check codes that cut qubit overhead for error correction by ~90 %, paving the way for modular machines like Starling (200 logical qubits) and Blue Jay  .

Google Willow: Willow scales error‑correction below threshold, halving errors as more qubits are added; it ran a random‑circuit sampling task in five minutes that would take a supercomputer 10^25 years  .

Reliable logical qubits: Quantinuum and Microsoft built four logical qubits with error rates 800× lower than the physical qubits, then entangled 12 logical qubits in a GHZ state with circuit errors of 0.0011  .

50‑qubit entanglement: Quantinuum later entangled 50 logical qubits with >98 % fidelity and used single‑shot error‑correction techniques.

Infleqtion: Neutral‑atom pioneer Infleqtion announced a 1 600‑qubit array and aims to deliver a 100‑logical‑qubit, million‑depth circuit machine within five years  .

IonQ’s scale-up: IonQ plans 100 qubits in 2025, 10 000 on one chip by 2027, and 2 million by 2030, translating to tens of thousands of logical qubits with error rates <10^‑12  .

Algorithmic fault tolerance: Harvard, Yale and QuEra researchers unveiled a framework that combines transversal operations with correlated decoding, promising 10–100× faster execution of large‑scale algorithms  .

Exponential speedup: A USC‑led team used IBM’s 127‑qubit chips to demonstrate an unconditional exponential speedup on a variation of Simon’s problem by combining shorter circuits, dynamical decoupling and error‑mitigation  .

Industrial silicon qubits: Imec and Diraq showed that CMOS‑fabbed silicon quantum‑dot qubits achieve >99 % two‑qubit gate fidelity and >99.9 % SPAM fidelity, making high‑volume manufacturing feasible .

Qudit error correction: Yale researchers demonstrated error‑corrected qutrits and ququarts using the GKP bosonic code, pushing these higher‑dimensional qudits past break‑even